DEV Community: neo one

AI Bookkeeping for Solopreneurs: Run a Clean Financial Stack Without an Accountant

neo one — Fri, 20 Mar 2026 13:10:07 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

AI Bookkeeping for Solopreneurs: Run a Clean Financial Stack Without an Accountant

Running a one-person business means you handle everything. Including the part everyone hates.

Finance. Bookkeeping. Taxes. The work that has nothing to do with why you started the business and everything to do with whether it survives.

Most solopreneurs automate writing, research, customer support, and outreach before they ever touch their books. That is backwards. Finance errors compound. A missed expense category costs you at tax time. An untracked invoice costs you cash. A messy P&L costs you decisions.

Automate finance first. Everything else is downstream of whether you know your numbers.

This guide covers the tools, the setup, and what to hand off first.

What You Actually Need

Four things cover a complete financial operation for a solopreneur:

A bank that helps you see the numbers. Not just hold the money.
Bookkeeping that does not require you to understand double-entry accounting.
Invoicing that actually gets paid.
Tax prep that does not take a week of your life.

Here is what covers each layer.

Banking: Mercury or Relay

Mercury is the default choice for solopreneurs and indie founders. It is built for exactly this use case.

What it does beyond holding money:

Organized transactions with automatic categorization
Virtual cards for vendor tracking
Automated savings rules (set aside 30% of every deposit for taxes, automatically)
Treasury account for operating reserves
API access if you want to pipe data to other tools

Relay is the alternative if you want multi-account structures. Good for keeping tax reserves, operating funds, and savings in separate accounts rather than tracking by rule. The visibility is cleaner if you think in buckets.

Both have no monthly fees. Both are built for small businesses. Neither is a traditional bank.

Skip traditional banks for your business account. They are not built for this.

Bookkeeping: Digits or Bench

Digits

Digits is AI-native bookkeeping. Connect your bank accounts and it categorizes transactions, tracks your financial health, and generates the reports you actually need.

This is not a human service. It is software that reads your finances and surfaces what matters. The reporting is fast. The insight layer is genuinely useful. You get a real-time view of your business without doing any of the work.

Good for: founders who want clean books and fast visibility without hiring anyone.

Bench

Bench is a hybrid. You get software plus a human bookkeeper assigned to your account. They reconcile monthly, clean up edge cases, and handle your books end to end. Bench uses software to do the heavy lifting and humans to catch what the software misses.

Good for: founders who want a professional to handle it and would rather not look at a dashboard at all.

Which one?

If your transactions are clean and predictable, use Digits. If you have messy books, mixed business and personal accounts, or irregular income, start with Bench and clean it up over a few months.

Either way, you need one of them. Running books manually in a spreadsheet is not a financial strategy. It is a way to lose money slowly.

Expense Tracking: Ramp or Brex

Business cards with built-in expense intelligence. This category changes how much time you spend on bookkeeping more than any other tool.

Ramp automatically categorizes every purchase, enforces spending rules, and flags anomalies. You get an expense report without building one. AI matches receipts to transactions. Accounting integrates directly with QuickBooks or Xero if you use those.

Brex does the same thing with a heavier tilt toward venture-backed startups, but accessible to solopreneurs. Better rewards if you spend heavily on software subscriptions.

Either card eliminates manual expense reports. That alone is worth it. Put every business expense on one of these and you will never manually reconcile a credit card statement again.

Invoicing: Stripe or FreshBooks

Stripe

If you already use Stripe for payments, use Stripe Invoicing. It handles one-time invoices, recurring subscriptions, automated payment reminders, and revenue recognition in one place.

The move most solopreneurs miss: automate payment reminders. Three reminders before the due date, one on the due date, one after. Most late payments are not intentional. They are forgotten. Automated reminders collect 80% of those without any back and forth.

FreshBooks

Better if you are a service provider billing by project or time. FreshBooks was built for freelancers and handles proposals, contracts, and invoicing in one product. The interface is simpler than QuickBooks and more appropriate for a one-person operation.

Use Stripe if your business is product or subscription-based. Use FreshBooks if you invoice clients for services.

Receipt Management: Dext

Dext (formerly Receipt Bank) captures receipts and feeds them into your bookkeeping software automatically.

The workflow: take a photo of a receipt. Dext reads the vendor, amount, and date. It categorizes the expense. It syncs to Bench, Digits, QuickBooks, or Xero.

This sounds minor. It is not. The hours spent hunting for receipts at tax time are not fun hours. Dext eliminates them.

Set it up. Use it for every paper receipt. The digital ones get captured automatically via email forward.

Tax Prep: What to Automate, What to Hand Off

You can automate the inputs. You cannot automate the judgment calls.

Automate:

Quarterly estimated tax transfers. Mercury's automated savings rules move a percentage of every deposit to a tax account. You never have to remember to set money aside.
Receipt capture. Dext handles this automatically once set up.
Expense categorization. Your bookkeeping software handles this.
Annual P&L and balance sheet export. Digits or Bench generates these in one click.

Do not automate:

Filing. Use a CPA who handles small business returns. It costs $500 to $1,500 per year. It saves you more than that in deductions you did not know to take.
Deduction decisions. A CPA knows what is defensible. Software makes assumptions. Your specific situation requires judgment.

The goal is to arrive at tax season with clean books, organized receipts, and a complete P&L ready to hand over. A good CPA takes it from there in a few hours instead of a few weeks.

Using AI for Financial Analysis

Once you have clean data in Digits or Bench, use AI to analyze it.

Practical ways to use Claude or any capable LLM with your financial data:

Drop your monthly P&L in and ask: what are my three highest-cost categories and are they trending up or down?
Compare two months: where did I spend more in March than February and what might explain it?
Forecast: based on my last six months, what is a realistic Q3 revenue target if I add one new client?
Identify waste: are there recurring charges I should audit or cancel?

This is not magic. It requires clean data going in. If your books are a mess, the analysis is useless. Clean books first. AI analysis second.

For founders who want this more automated, Make.com can export your monthly financial summary and send it to Claude via an API call. You get a plain-English financial briefing in your inbox on the first of every month.

Wiring It Together with Automation

Make.com or n8n can connect your finance tools into automated workflows:

New Stripe payment received. Create a record in Notion. Update your revenue tracker.
Expense over $500. Flag in Slack. Require a category tag before it closes.
Monthly close complete. Export P&L from Digits. Send summary to a Notion dashboard.
30 days before quarterly tax date. Remind you to check your estimated payment amount.
Invoice 7 days overdue. Trigger a follow-up email automatically.

None of these require AI. They are deterministic automation. Set them up once and they run indefinitely.

Use agents where rules break down. Use automation everywhere else.

The Minimal Setup (Start Here)

If you are starting from scratch or your current setup is a mess:

Open a Mercury account. Move all business money there. This is your financial foundation.
Get Ramp. Use it for every business expense. No exceptions.
Set up Stripe Invoicing if you invoice clients. Enable automated payment reminders.
Connect Digits or open a Bench account. Do this in the first week.
Set a Mercury rule to move 30% of every deposit to a tax savings account automatically.
Download Dext. Forward all receipt emails to your Dext inbox. Photograph every paper receipt.
Find a CPA who handles sole proprietors. One conversation in January. They handle the rest.

That is a complete financial stack. Total cost: under $100 per month. It replaces hours of manual work. It gives you accurate numbers. It means you do not dread tax season.

What This Actually Unlocks

You will know if your business is working. Not at year end. Every week.

Most solopreneurs do not know if they are profitable until they file taxes. That is not a financial problem. That is an information problem.

You cannot make good decisions without good information. You cannot raise prices confidently if you do not know your margins. You cannot plan hires or tool purchases if you do not know your runway.

Solve the information problem first. The rest gets easier.

Related Guides

Top No-Code and Low-Code Agent Frameworks for Non-Engineers

neo one — Fri, 20 Mar 2026 13:10:01 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

Top No-Code and Low-Code Agent Frameworks for Non-Engineers

Not everyone who needs an AI agent workflow wants to spend a weekend debugging Python dependencies. Valid.

This guide covers the leading no-code and low-code options for building agent workflows, what they are actually good at, where their limits are, and when you will need to bring in an engineer anyway. Because that moment does come, and knowing it in advance saves you from building the wrong thing.

The good news: you can accomplish more than you might think without code. The honest caveat: the most powerful agent systems still require someone who can write code, debug APIs, and handle edge cases. But the no-code tier has gotten dramatically more capable in the last 18 months.

How to Think About This Category

No-code agent frameworks exist on a spectrum:

Pure no-code: Visual workflow builders where everything is drag-and-drop. Fast to start, limited in ceiling.

Low-code: Mostly visual, but with the option to write expressions, scripts, or SQL when needed. Better ceiling, slightly steeper learning curve.

Code-first: Full SDKs and frameworks (LangChain, LlamaIndex, AutoGen). Highest power, requires engineering skills. Not covered in this guide.

For most non-engineers, the right tool is in the low-code tier, enough power to handle real business workflows, without needing to understand HTTP headers or Python environments.

The Tools

n8n

Best for: Technical non-engineers who want serious automation power without writing full applications

n8n is a workflow automation platform with deep AI agent support. You build workflows visually, connecting nodes that represent actions (HTTP calls, database queries, LLM prompts, API integrations). When you need custom logic, you can add JavaScript snippets in specific nodes.

What it does well:

400+ native integrations
AI Agent node that handles tool-calling, memory, and multi-step reasoning natively
Can be self-hosted (important for data-sensitive businesses)
Sub-workflows that let you build modular, reusable agent behaviors
Strong error handling and retry logic
Active open-source community with a library of workflow templates

Limitations:

Has a learning curve, the visual interface isn't as intuitive as Zapier for simple tasks
Debug experience is clunky for complex workflows
Self-hosting requires some infrastructure knowledge

Pricing: Free self-hosted. Cloud plans from $20/month.

Who it's for: Ops leads, technical founders, and power users comfortable with logic but not full software engineering.

Make (formerly Integromat)

Best for: Complex multi-step workflows with good balance of power and usability

Make is the serious alternative to Zapier for people who've hit Zapier's ceiling. Its visual workflow builder shows the data flowing between steps, which makes it much easier to understand complex pipelines.

What it does well:

Advanced routing and filtering logic
Iteration over arrays and data structures
HTTP modules for calling any API
Native OpenAI and AI model integrations
Good error handling with auto-retry
Significantly cheaper than Zapier at volume

Limitations:

Less polished UI than Zapier
Native AI Agent support is less mature than n8n's dedicated agent node
Can get expensive at high operation volume

Pricing: Free plan (1,000 ops/month). Paid from $9/month.

Who it's for: Anyone who needs more than Zapier can do but isn't ready to self-host n8n.

Zapier

Best for: Simple, reliable automations without any technical friction

Zapier is the most-used automation platform for a reason: it's extremely easy to set up, has the widest app library (7,000+ apps), and has a polished interface that non-technical users can navigate without training.

What it does well:

Best-in-class ease of use and onboarding
Reliable for straightforward trigger-action workflows
Zapier AI features (Copilot, AI steps) are improving
No learning curve for basic automations

Limitations:

Expensive at volume (pricing is per task, and it adds up fast)
Limited branching logic and data transformation
The ceiling for complex agent behavior is low compared to n8n or Make
AI agent capabilities are nascent, it can call LLMs, but true agent loops are not well-supported

Pricing: Free plan (100 tasks/month). Paid from $19.99/month. Gets expensive fast.

Who it's for: Non-technical users who need simple automation and can afford the convenience premium.

Relevance AI

Best for: Building AI agents specifically, without worrying about general workflow automation

Relevance AI is purpose-built for AI agents. Instead of starting with workflows, you start with agent behaviors. You define what your agent knows, what tools it has, and what goals it pursues, all through a no-code interface.

What it does well:

Agent-first design, feels natural for AI agent use cases
Tool builder lets you create custom tools without code (wraps APIs in an agent-friendly interface)
Multi-agent orchestration: create teams of agents that hand off to each other
Memory and knowledge base management for agents
Good for business users building internal AI assistants or sales intelligence tools

Limitations:

Less flexible than n8n for general-purpose automation
Newer platform with a smaller community
More expensive at scale than general automation tools

Pricing: Free trial. Paid plans from $19/month per agent.

Who it's for: Business users who want to build agents (not just automations) and prefer an interface designed around AI from the start.

Voiceflow

Best for: Building conversational AI agents (chatbots, voice assistants, support agents)

Voiceflow is the best tool in its category for designing conversational experiences. If your use case involves an AI that talks to users, customer support, lead qualification, onboarding flows, Voiceflow gives you a visual conversation designer that's more powerful than building in a raw LLM interface.

What it does well:

Visual conversation flow design
Multi-channel deployment (web chat, Slack, WhatsApp, Zendesk, voice)
Variable handling and conditional branching
Integration with CRMs and support platforms
Good for non-developers building internal or customer-facing chat experiences

Limitations:

Primarily for conversational agents, not process automation
More complex data processing requires workarounds
Cost scales with message volume

Pricing: Free for limited use. Team plans from $50/month.

Who it's for: Customer success teams, support leads, and product managers building chatbot or voice agent experiences.

Flowise

Best for: Self-hosted, open-source AI workflow building with a visual interface

Flowise is an open-source alternative that wraps LangChain in a drag-and-drop interface. It's free to self-host and gives non-developers access to LangChain's power without writing Python.

What it does well:

Free and self-hostable
Strong support for RAG (retrieval-augmented generation) workflows
Connects to most major LLM providers
Active open-source community

Limitations:

Still requires basic technical comfort to deploy and maintain
Less polished than commercial alternatives
Debug experience is rough

Pricing: Free (self-hosted). Cloud version available.

Who it's for: Technical operators who want open-source control and are willing to manage their own deployment.

Comparison at a Glance

Tool	Best For	Ease of Use	Agent Capability	Price Entry
n8n	Power automation + agents	Medium	High	Free (self-hosted)
Make	Complex workflows	Medium	Medium	$9/mo
Zapier	Simple automations	High	Low	$20/mo
Relevance AI	Agent-first building	High	High	$19/mo
Voiceflow	Conversational agents	High	Medium (conversational)	$50/mo
Flowise	Open-source RAG agents	Low-Medium	High	Free

How to Choose

Start with Zapier if: You need 1-3 simple automations, have no technical background, and the time savings justify the cost.

Start with Make if: You need multi-step workflows with data transformation and want more power than Zapier at a lower price.

Start with n8n if: You want maximum power and flexibility, you're comfortable with some technical complexity, and you care about self-hosting or data sovereignty.

Start with Relevance AI if: Your primary goal is building AI agents (not general automation) and you want an interface designed specifically for that.

Start with Voiceflow if: You're building a conversational AI that talks to users, customer support, lead qualification, onboarding.

Start with Flowise if: You want LangChain power without code and are willing to self-host.

When You'll Need an Engineer Anyway

No-code tools have real ceilings. Here's when to escalate:

Custom authentication: OAuth flows, API key rotation, and non-standard auth schemes often require code
Real-time data processing: Streaming, WebSockets, and low-latency requirements push past what visual tools handle well
Complex error handling: When workflows need to handle partial failures, retries with backoff, and edge cases gracefully
Data transformation at scale: Processing large datasets efficiently requires code
Production reliability: Mission-critical workflows need monitoring, alerting, and CI/CD that visual tools don't support well
Custom tool development: Building new agent tools (not wrapping existing APIs) requires code

For many business workflows, automating data entry, routing information between systems, triggering LLM tasks on events, you won't hit these limits. But know they exist.

Getting Started Recommendation

If you're new to this space, pick one tool and build one workflow from start to finish before evaluating alternatives. The hands-on experience will teach you more than any comparison guide.

A good first workflow to build: "When a form is submitted, enrich the lead data, draft a personalized follow-up email using an LLM, and send it for human review before it goes out." This touches triggers, API calls, LLM prompting, and human-in-the-loop, the core components of most AI agent workflows.

Related Guides

n8n vs Make for AI Agents, Deep comparison of the two leading low-code platforms
How to Delegate Tasks to AI Agents, Designing agent tasks that actually work
MCP Tool Ecosystems, How to connect agents to tools at a protocol level
Getting Started with AI Agents, Core agent concepts before choosing a framework

The AI Hiring Stack: Tools for Recruiting with AI Agents

neo one — Fri, 20 Mar 2026 13:04:50 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

The AI Hiring Stack: Tools for Recruiting with AI Agents

Hiring is the job that never ends and the one everyone does wrong. You post a role, get 200 applications, spend 12 hours screening, interview six people, hire one, and repeat in three months.

AI does not fix the repeat. But it handles most of the other steps. This guide covers the full recruiting funnel: sourcing, screening, interview prep, and offer management.

If you're running lean or hiring for the first time, this guide will help you avoid the common mistakes and identify the highest-leverage tools available today.

Where AI Actually Helps in Recruiting

Before picking tools, understand where AI delivers real leverage versus where it adds noise.

High-leverage uses:

Writing and iterating on job descriptions
Parsing and scoring large applicant volumes
Drafting interview questions tailored to the role
Summarizing interview notes
Generating rejection and offer email templates
Researching candidate backgrounds quickly

Low-leverage or risky uses:

Final hiring decisions (bias risk, legal liability)
Personality or culture-fit assessment
Reference checks (relationships require humans)
Salary negotiation (too relationship-dependent)

The clearest ROI comes from reducing time-on-task for repetitive steps, writing, scoring, scheduling, so your team spends more time on the parts that actually require human judgment.

The Five Stages of the AI Recruiting Funnel

Stage 1: Writing Job Descriptions

Most job descriptions are either too vague ("seeking a rockstar") or too prescriptive (17 requirements for a mid-level role). Both kill your candidate pipeline.

What AI does well here: Claude or GPT-4 can take a bullet list of responsibilities and produce a structured job description in under a minute. The more specific your input, the better the output.

Prompt template that works:

Role: [title]
Team: [who they'll work with]
Responsibilities: [bullet list]
Must-have qualifications: [list]
Nice-to-have: [list]
Salary range: [if sharing]
Company context: [2-3 sentences on stage/product]

Write a job description that's specific, honest, and avoids filler phrases. Target candidates who are early-career to mid-level. Keep it under 500 words.

Iterate once or twice. The final output will be better than most human-written JDs in half the time.

Tools: Claude, ChatGPT, or Notion AI if you draft in Notion.

Stage 2: Sourcing Candidates

Sourcing is one of the most time-intensive parts of recruiting. AI agents are starting to change this meaningfully.

LinkedIn + AI outreach tools:

Waalaxy and Dripify, automate LinkedIn connection sequences with personalized message templates. You write the messaging logic; they execute it at scale.
Clay, the most powerful tool in this category. Clay lets you build lead lists of candidates from LinkedIn, GitHub, Twitter, or company databases, enrich them with data (title, company size, open-source contributions), and draft personalized outreach using AI. Originally built for sales, increasingly used for recruiting.
Gem, designed specifically for recruiting, integrates with your ATS, and uses AI to surface past candidates who might fit new roles.

GitHub sourcing: For engineering roles, GitHub is often more signal-dense than LinkedIn. Tools like GitClear or GitHub's own search let you identify contributors to relevant open-source projects. You can feed those profiles into Clay to enrich and message.

The agent approach: If you're building your own stack, you can build a sourcing agent using n8n or Make that pulls candidates from a LinkedIn search URL, enriches them via Clay or Apollo, scores them against your criteria using an LLM, and pushes qualified candidates to a review sheet for human approval before any outreach goes out.

Stage 3: Screening Applications

For high-volume roles, application screening is where AI pays the fastest dividend.

Resume parsing and scoring: Tools like Ashby, Greenhouse, and Lever have AI-assisted screening built in. They can rank applicants against role criteria and surface the top N for human review.

Custom scoring with AI: If you want more control, you can build your own scoring pipeline. Export applications from your ATS to a spreadsheet or Airtable. Use an AI prompt to score each application against your specific criteria (not just keywords). Output a score and a one-paragraph rationale for each.

Example scoring prompt:

Role criteria:
- Must have shipped a product with real users
- Experience with Postgres at scale
- Has worked on a small team (< 10 engineers)

Here is the candidate's resume: [resume text]

Score 1-10 on each criterion. Give a one-sentence rationale for each score. Sum the scores and give a final recommendation: Strong Yes, Yes, No.

Caveat: AI screening should narrow, not decide. Present top candidates to a human for final shortlisting. Document your scoring criteria to reduce bias exposure.

Pre-screening assessments: Tools like Karat (technical interviewing), Codility, and HackerRank run skill-based assessments before interviews. The results are objective and easy to compare.

Stage 4: Interviews

AI won't run your interviews. But it can make them dramatically better.

Interview prep with AI:

Input the job description and the candidate's resume.
Ask the model to generate 10 role-specific behavioral and technical questions.
Ask it to identify 2-3 areas of the resume worth probing (gaps, ambiguous scope, fast transitions).
Ask it to draft the scorecard criteria you'll use to evaluate answers.

This takes 5 minutes and produces a structured interview that's more consistent across candidates.

Transcription and note summarization: Tools like Otter.ai, Fireflies, and Fathom transcribe interviews and can produce summaries. For recorded async video interviews, Hireflix or Spark Hire let candidates record answers to your questions, which you review asynchronously, no scheduling required for the first round.

Interview intelligence platforms: Greenhouse and Ashby track which interview questions correlate with hiring decisions and downstream performance. If you're running enough volume, this data becomes actionable.

Stage 5: Offer Management and Closing

Offer management is typically manual, email-heavy, and slow. AI can help with drafting and structuring, but closing still requires relationship.

Offer letter drafts: Give your LLM the offer details (title, salary, equity, start date, benefits summary) and ask it to draft a warm, specific offer letter. The output needs legal review for your jurisdiction, but the draft saves an hour.

Offer communication templates: Prepare templates for:

Initial offer email (warm, specific to the candidate)
Follow-up if no response in 48 hours
Negotiation response (acknowledging counter, holding or adjusting)
Acceptance confirmation with next steps
Decline response (keeping the relationship warm)

Closing conversations: This is where AI stops helping. Compensation negotiation and final-stage persuasion require knowing the candidate's real situation, family constraints, competing offers, their specific hesitations. These conversations need a human who has built rapport.

The Lean AI Hiring Stack (For Solo Operators and Small Teams)

If you're hiring 1-5 people per year:

Stage	Tool	Cost
Job description	Claude or ChatGPT	Free/low
Job posting	LinkedIn, Indeed	Per posting
Application tracking	Notion or Airtable	Free tier
Screening	AI prompt on spreadsheet	Free
Interviews	Fathom (transcription)	Free tier
Offers	LLM-drafted, manual send	Free

Total tooling cost for lean hiring: near zero. The leverage is in using AI for the writing and scoring work.

The Growth Stack (For Teams Hiring at Volume)

If you're hiring 20+ people per year:

Stage	Tool	Why
Sourcing	Clay + LinkedIn Recruiter	Best enrichment and outreach
ATS	Ashby or Greenhouse	AI features built in
Screening	Codility or Karat (technical)	Objective skill data
Interviews	Fathom + structured scorecards	Consistency at scale
Analytics	Greenhouse reporting	Track pipeline health

What AI Cannot Replace in Hiring

Judgment about culture and values, only humans who know your team can assess this
Candidate experience, fast, warm, human communication is a competitive advantage
Closing strong candidates, top candidates have options; closing requires real conversation
Legal compliance, AI tools don't know your local labor laws; run all AI-generated content through a people ops or legal review

Related Guides

AI Agents for Solopreneurs, How agents work before you apply them to HR
How to Delegate Tasks to AI Agents, Delegation patterns that apply to recruiting workflows
Solo AI Business Infrastructure, The broader stack for running an AI-forward business

How to Run a Content Operation with AI Agents End-to-End

neo one — Fri, 20 Mar 2026 13:04:42 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

How to Run a Content Operation with AI Agents End-to-End

Running a content operation used to mean hiring people. An editor. A few writers. Someone to handle social. Someone who actually publishes it without forgetting the meta description. In 2026, most of those roles are software.

This guide covers the entire pipeline, from identifying what to write through publishing and distribution, and maps exactly where AI adds leverage versus where a human still needs to be in the room.

A well-designed AI content operation can produce more output, more consistently, at lower cost. It can also produce a lot of generic noise if you're not careful. The difference is in how you design the pipeline.

What a Content Operation Actually Is

A content operation is not just "writing blog posts." It's a system with inputs, processes, and outputs that compounds over time.

The stages:

Strategy and ideation, What to write and why
Research, What's true, what's current, what the audience needs
Writing, First draft production
Editing, Quality control, voice, accuracy
Publishing, Getting it into the CMS with proper metadata
Distribution, Getting it in front of the right people
Performance tracking, What's working and what to do more of

AI can touch every stage. The question is: how much, and with what human review?

Stage 1: Strategy and Ideation

Content strategy requires understanding your audience, your positioning, and your competition. This is where human judgment matters most, but AI can do the legwork.

What AI can do:

Pull keyword and topic data from Ahrefs, Semrush, or Search Console
Summarize what competitors are publishing
Generate topic lists from a brief or a list of customer questions
Score topic ideas against traffic potential, difficulty, and relevance

What humans need to do:

Decide which topics align with your positioning
Prioritize based on business goals (not just traffic)
Identify the angle that differentiates your content

Tool setup: Use Ahrefs or Semrush for data. Feed keyword clusters to Claude or ChatGPT and ask it to group them into content themes and suggest 10 specific article angles per theme. You review and curate. The AI generates the raw material; you apply editorial judgment.

Content brief generation: Once you pick a topic, use an AI prompt to generate a structured brief: target keyword, search intent, key questions to answer, sections to include, related topics to mention. A good brief takes 2 minutes to generate and 5 minutes to review. It makes the writing stage much faster.

Stage 2: Research

Research is the stage where most AI content fails. Models trained on data have a cutoff date and hallucinate specifics. If your content requires current statistics, product feature details, or verified claims, you cannot rely on the model's training data.

Research approaches that work:

Web search-augmented generation: Tools like Perplexity, You.com, or Claude with web access can pull current information and cite sources. Use these when currency matters.

Source-first research: Collect your sources manually (or via a research agent), then ask the model to synthesize them. This is the highest-quality approach.

SERP analysis: Ask an AI to analyze the top 10 results for a given keyword and identify what they all cover, what they miss, and what questions they leave unanswered. This produces a gap analysis that is genuinely useful for angle differentiation.

Expert quote collection: For authoritative content, scrape or search for quotes from known practitioners on the topic. Feed those quotes to an AI and ask it to identify the most relevant ones and draft context around them.

Research agent workflow (n8n or Make):

Trigger: content brief finalized
Pull top 10 SERP results for target keyword
Scrape and extract main body text from each
Send to LLM for gap analysis and key point extraction
Output: research brief with sources, key angles, existing coverage gaps

This automates 30-40 minutes of research work per article.

Stage 3: Writing

This is where most people start, and where the most nuance lives.

The failure mode: Prompting "write me a 2000-word article on X" and publishing what comes out. This produces generic content that ranks poorly and reads like it was written by a committee.

The approach that works: AI writes the structure and fills sections; a human with domain knowledge adds the differentiation.

Draft generation workflow:

Start with the research brief
Generate a section-by-section outline, reviewed by a human
Generate each section individually with specific prompts
Paste sections together into a single document
Human review pass: add specific examples, fix voice, cut filler

Generate sections individually rather than the whole article at once. You get more control, better quality, and it's easier to catch errors.

Voice and style: This is the biggest gap in AI-written content. To close it:

Create a style guide document and paste it into every writing prompt
Give the model 3-5 examples of your best existing content and ask it to match the style
Be explicit about what you don't want: no clichés, no hedging, no "it's important to note that"

Specialist tools for writing:

Jasper, built for marketing copy, has brand voice features
Copy.ai, strong for short-form and structured templates
Claude, best for long-form and research-heavy content
Notion AI, if your team already drafts in Notion

Stage 4: Editing

Editing is where AI content goes from mediocre to good. Do not skip this stage.

What AI can help with:

Proofreading (Grammarly, LanguageTool)
Readability scoring and suggestions (Hemingway Editor)
Identifying repetition and filler phrases
Fact-checking against provided sources
SEO metadata (meta descriptions, title tags) generation

What humans must own:

Voice consistency, does this sound like us?
Accuracy of specific claims, did the AI get this right?
Angle differentiation, does this say something the competition doesn't?
Legal and compliance review, for regulated industries

Editing prompt template:

Review this draft for:
1. Filler phrases to cut (list them)
2. Sections that are vague where a specific example would be stronger
3. Any claims that need a source or caveat
4. The opening paragraph, rewrite it to hook faster

Do not change the structure or main points. Just flag and suggest.

This produces a structured edit report faster than reading the whole draft fresh.

Stage 5: Publishing

Publishing is highly automatable and often the most time-wasting manual step in a content operation.

CMS integrations:

Contentful, Sanity, Ghost, and WordPress all have APIs that AI agents can write to
Zapier and Make have native connectors for most CMS platforms

Automated publishing workflow:

Final draft approved in a shared doc (Notion, Google Docs)
Webhook or manual trigger starts the pipeline
Agent parses the document and extracts: title, body, meta description, tags, featured image prompt
Image generated via DALL-E, Midjourney API, or Replicate if needed
Content pushed to CMS via API
URL added to a tracking spreadsheet

This workflow takes what used to be 20-30 minutes of manual CMS work and reduces it to a review-and-approve step.

SEO metadata automation: When generating metadata, always include the target keyword in a specific prompt:

Write an SEO meta description for this article:
- Target keyword: [keyword]
- Article topic: [topic]
- Key benefit for the reader: [benefit]
- Max 155 characters
- Do not start with "This article" or "In this guide"

Stage 6: Distribution

Publishing is not distributing. Most content fails because it's published and forgotten.

Distribution assets AI can generate from a single article:

3-5 LinkedIn posts with different angles
5-10 tweets or X posts
Email newsletter section (summary + link)
Short-form video script for TikTok or Reels
A thread format for X or Reddit
A podcast episode outline

The repurposing prompt:

Here is a [topic] article: [paste article]

Generate:
1. A LinkedIn post (400-600 words, first-person, starts with a hook, no hashtag spam)
2. A 5-tweet thread that covers the main points
3. A 150-word email newsletter paragraph that teases the article and links out

Style guidance: [paste style guide or examples]

This takes 2 minutes and produces 3 distribution assets. A human edits them lightly for voice, then schedules.

Distribution automation: Tools like Buffer, Hypefury, and Publer can schedule posts across platforms. Wire them to your content tracker so distribution is triggered automatically when an article publishes.

Email distribution: ConvertKit, Beehiiv, and Ghost all have API or Zapier integrations. You can automate sending a newsletter digest triggered by new published articles.

Stage 7: Performance Tracking

Content that isn't tracked doesn't improve.

What to track:

Organic traffic per article (Search Console, GA4)
Rankings for target keywords (Ahrefs, Semrush)
Email click-through from newsletter links
Social engagement per distribution asset
Leads or signups attributed to content

AI reporting: Feed your analytics data into an LLM weekly and ask it to:

Identify the top 3 performing articles and what they have in common
Flag articles with declining traffic that might need updates
Suggest 5 new article ideas based on current winners

This turns a manual data review into a 5-minute conversation.

The Lean AI Content Stack (Solo Operator)

Stage	Tool	Cost
Ideation	Claude + free Ahrefs tier	Free/low
Research	Perplexity + manual	Free
Writing	Claude	~$20/mo
Editing	Hemingway + manual	Free
Publishing	Ghost + manual API	$9/mo
Distribution	Buffer free tier	Free
Tracking	Search Console + GA4	Free

Monthly cost: ~$30. Output capability: 8-12 quality articles per month with 2-3 hours of human time per article.

The Growth Stack (Small Team)

Stage	Tool	Cost
Strategy	Ahrefs + Claude	$99/mo + $20/mo
Research	Perplexity Pro + scrapers	$20/mo
Writing	Claude + Jasper	$20-$49/mo
Editing	Grammarly Business	$15/mo
Publishing	Contentful + n8n automations	Varies
Distribution	Hypefury + Buffer	$49/mo
Tracking	GA4 + Ahrefs	Included

What to Never Fully Automate

Final editorial approval, someone with taste and judgment must read every piece before it publishes
Response to comments and community, genuine engagement requires genuine people
Expert opinions and original research, AI can summarize existing views, not create new ones
Sensitive topics, health, legal, financial content requires qualified human review

Related Guides

How to Delegate Tasks to AI Agents, Delegation patterns that apply to content workflows
Building a Personal AI Knowledge Base, How to give your content agents memory and context
MCP Tool Ecosystems, Connecting AI agents to your publishing and analytics tools
Top MCP Servers for Business Automation, Tools for automating the publishing pipeline

16 One-Man Business Ideas in the AI Era

neo one — Fri, 20 Mar 2026 12:59:28 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

16 One-Man Business Ideas in the AI Era

The best one-man business ideas in 2026 don't feel like one-man businesses.

They feel like companies. Because they operate like companies. One founder, a stack of AI agents, and enough infrastructure to handle what used to need six people.

This is the actual play. Not "start a blog." Not "do freelancing." Every idea on this list runs primarily on AI. If the same business worked just as well in 2019, it's not on the list.

For each idea: what it is, what the AI handles, what you actually do, and the tools you need. That last column matters. These are real stack requirements, not vague suggestions.

1. AI Research Boutique

What it is: A one-person research firm that delivers structured intelligence reports on any topic. Competitors, markets, regulations, due diligence. Clients get in-depth analysis without hiring an analyst.

What the AI does:

Searches the web and synthesizes findings across dozens of sources
Structures reports into consistent, readable formats
Identifies gaps and contradictions in the source material
Generates follow-on questions from initial findings

What you do:

Define the research brief and quality bar
Review and edit the final output
Manage client relationships and communication
Identify which questions AI cannot yet answer reliably

Tools: Claude Opus 4.6, Perplexity or Exa for search, Make.com for intake automation, Notion for report delivery.

Revenue: $300 to $2,500 per report. Retainer clients at $3,000 to $8,000 per month.

2. AI Outbound Sales Operation

What it is: A one-person outbound sales agency that builds and runs lead generation systems for B2B companies. You deploy the pipeline. The agents run it.

What the AI does:

Finds target accounts matching the client's ideal customer profile
Enriches contact data and identifies the right decision-maker
Personalizes email sequences at scale using signal-based context
Follows up automatically and routes hot leads to the client

What you do:

Set strategy and targeting criteria
Review and approve copy before it goes out
Handle edge cases and complex objection responses
Report results and optimize based on reply data

Tools: Clay for lead enrichment, Claude Haiku 4.5 for personalization at volume, Make.com for sequencing, Instantly for sending.

Revenue: $3,000 to $8,000 per month per client on retainer.

Related: AI-Powered Outbound Sales

3. AI Content Factory

What it is: A one-person content operation that produces SEO content at agency output levels. You handle strategy and editing. AI handles research and first drafts.

What the AI does:

Conducts keyword research and competitive gap analysis
Writes structured first drafts from briefs
Generates internal linking recommendations
Produces meta descriptions, image alt text, and social variants

What you do:

Develop the content strategy and editorial calendar
Edit for accuracy, voice, and quality before publish
Manage client relationships and reporting
Make judgment calls on which content to prioritize

Tools: Ahrefs for keyword research, Claude Sonnet 4.6 for drafting, Make.com for workflow automation, Notion for editorial tracking.

Revenue: $3,000 to $12,000 per month per client.

4. AI Customer Support Provider

What it is: A one-person support operation that handles tier-one customer inquiries for small businesses. You build the system. The agents handle the volume.

What the AI does:

Responds to common inquiries using the client's knowledge base
Routes complex issues to the human escalation path
Drafts responses to edge cases for human approval
Identifies recurring issues for the client to address upstream

What you do:

Build and maintain the knowledge base
Handle complex escalations
Monitor quality and flag output drift
Onboard new clients and customize the agent per product

Tools: Claude Haiku 4.5 for response generation, Intercom or Zendesk as the support layer, Supabase with pgvector for knowledge base storage, Make.com for escalation routing.

Revenue: $1,500 to $5,000 per month per client.

Related: AI Agents for Customer Support

5. Vibe Coding Studio

What it is: A one-person software studio that builds and ships SaaS tools using AI coding assistants. You ship products faster than traditional development teams. You sell them directly or through marketplaces.

What the AI does:

Writes 80 to 90 percent of the actual code
Debugs and explains what went wrong
Suggests architecture and refactoring improvements
Generates tests, documentation, and deployment configs

What you do:

Define what to build and for whom
Review the code for correctness and security
Make product decisions about features and scope
Handle marketing, distribution, and customer feedback

Tools: Claude Code or Cursor for building, Vercel for hosting, Supabase for database and auth, Stripe for payments.

Revenue: $20k to $100k for AppSumo lifetime deals. $5k to $30k MRR for subscription products that find traction.

Related: Vibe Coding Tools

6. AI Social Media Ghost-Operator

What it is: A one-person operation that manages LinkedIn, X, or Instagram for executives and founders. You produce a full month of content in hours. The founder approves. You schedule and publish.

What the AI does:

Generates post ideas based on the client's expertise and recent activity
Drafts posts in the client's voice
Repurposes long-form content into shorter formats
Proposes engagement responses to high-value comments

What you do:

Develop and maintain the client's brand voice document
Edit AI output to match the client's authentic tone
Manage the approval and scheduling workflow
Track performance and adjust the content strategy

Tools: Claude Sonnet 4.6 for drafting, Buffer or Hypefury for scheduling, Make.com for the approval workflow.

Revenue: $1,500 to $5,000 per month per client.

Related: Delegating Creative Work to AI

7. AI Financial Reporting Service

What it is: A one-person fractional CFO service for early-stage companies. You take raw financial data and produce structured reports. Board decks. Cash flow summaries. Runway projections.

What the AI does:

Ingests raw data from Stripe, QuickBooks, or bank exports
Identifies anomalies and patterns in the numbers
Drafts narrative summaries explaining what the numbers mean
Generates charts and visualizations on demand

What you do:

Review the financial analysis for accuracy
Add context that requires knowledge of the specific business
Present findings to founders or boards
Flag risks and make recommendations

Tools: Claude Opus 4.6 for financial reasoning, Python or Make.com for data ingestion, Notion or Google Slides for report delivery.

Revenue: $1,000 to $5,000 per month per client.

8. AI Contract Review Service

What it is: A first-pass document review service for founders and small businesses. You flag unusual clauses, identify missing protections, and summarize what a contract actually says. Not legal advice. Intelligence before the lawyer.

What the AI does:

Reads and parses the full document
Flags clauses that deviate from standard terms
Summarizes each section in plain language
Suggests questions for the client to raise with legal counsel

What you do:

Define the review framework for each document type
Verify accuracy of AI output before delivery
Handle client questions and scope decisions
Build and maintain the library of what to look for by industry

Tools: Claude Opus 4.6, Make.com for file intake, Notion or PDF generation for delivery.

Revenue: $200 to $800 per document. $1,000 to $4,000 per month for clients with regular document flow.

9. AI Recruiting Pipeline

What it is: A one-person hiring operation that screens candidates and runs outbound sourcing for companies that hire frequently. You replace the function, not the judgment.

What the AI does:

Screens inbound resumes against defined criteria
Drafts personalized outreach to passive candidates
Sends initial screening questions and processes responses
Summarizes candidate pools for hiring manager review

What you do:

Define the hiring criteria with the client
Handle final-stage candidate evaluation
Manage the client relationship and intake process
Customize the screening framework per role

Tools: Claude Sonnet 4.6 for resume evaluation and outreach drafting, Make.com for ATS integration, Apollo for candidate sourcing.

Revenue: $3,000 to $10,000 per month per client.

Related: AI Hiring Stack

10. AI Monitoring and Intelligence Service

What it is: A signal-watching business. You deploy agents that monitor competitor moves, news mentions, regulatory changes, and market activity. Clients get a weekly intelligence brief instead of spending hours producing one themselves.

What the AI does:

Monitors specified sources on a defined schedule
Classifies and ranks signals by relevance to the client
Summarizes findings in a structured brief
Flags urgent signals for immediate notification

What you do:

Define the monitoring scope and relevance criteria
Review and edit the weekly brief
Update the system when the signal landscape shifts
Manage the client relationship and delivery

Tools: Exa or Perplexity for web monitoring, Claude Haiku 4.5 for classification and summarization, Make.com for scheduling, Slack or email for delivery.

Revenue: $500 to $3,000 per month per client. High margin once the system is built.

11. AI Podcast Production Studio

What it is: A one-person production house for podcasters and video creators who want more output with less effort. You handle the post-production pipeline. AI does most of the repetitive work.

What the AI does:

Transcribes episodes and generates timestamped summaries
Drafts show notes, titles, and episode descriptions
Identifies clip-worthy moments for social repurposing
Generates chapter markers and searchable episode indexes

What you do:

Edit the actual audio or video
Review AI-generated text before publishing
Manage client relationships and feedback
Keep the production calendar on schedule

Tools: Claude for scripting and show notes, Descript or Otter.ai for transcription, Opus Clip for repurposing.

Revenue: $2,000 to $6,000 per month per client.

12. AI Newsletter Business

What it is: A solo-run niche newsletter that delivers genuine intelligence to a targeted audience. Not a content dump. A curated, synthesized view of a specific topic. Monetized with sponsorships, a paid tier, or an adjacent product.

What the AI does:

Monitors the topic landscape and surfaces what is worth covering
Synthesizes multiple sources into a structured weekly view
Generates draft editions from a structured brief
Tests subject lines and formats

What you do:

Develop the editorial perspective that AI cannot replicate
Edit every edition before it goes out
Cultivate the audience and manage growth
Sell and manage sponsorships

Tools: Claude Sonnet 4.6 for drafting, Beehiiv or Resend for delivery, Make.com for monitoring and curation pipelines.

Revenue: $2,000 to $20,000 per month depending on audience size and monetization model.

13. Vertical AI SaaS Product

What it is: A narrow software tool built for one specific profession. Real estate agents. Dentists. Lawyers. Accountants. The tool handles the most repetitive, time-consuming task in their workflow. You sell it as a subscription.

What the AI does:

Performs the core computation the user is paying for
Handles onboarding and in-app guidance
Triages support questions and drafts help responses
Monitors usage patterns that predict churn

What you do:

Define the problem to solve and validate it with real users
Build and maintain the product
Acquire and retain customers
Make product decisions about what to build next

Tools: Claude API with Haiku 4.5 for cost efficiency at scale, Vercel, Supabase, Stripe.

Revenue: $99 to $499 per month per seat. At 100 customers that is $10,000 to $50,000 MRR.

Related: AI Solopreneur Business Models

14. AI E-commerce Operations Service

What it is: A one-person operations service for e-commerce brands. You produce product listings, ad copy, and email flows using AI. Clients measure the output in conversion rates. The ROI is visible.

What the AI does:

Generates product listing variations for A/B testing
Writes ad copy in multiple formats for Meta, Google, and TikTok
Drafts email sequences for cart abandonment, post-purchase, and win-back
Analyzes performance data and recommends copy adjustments

What you do:

Set the strategy and positioning for each client
Review and approve copy before it runs
Interpret performance data and make strategic calls
Manage the client relationship and reporting

Tools: Claude Sonnet 4.6 for copywriting, Klaviyo for email, Meta and Google for ad deployment.

Revenue: $3,000 to $12,000 per month per client with three to six clients.

15. AI Due Diligence Service

What it is: A research and analysis service for investors, acquirers, and executives who need to evaluate companies or markets quickly. You produce structured due diligence memos faster than any traditional research firm.

What the AI does:

Gathers and synthesizes publicly available information on a target company
Identifies red flags in disclosures, news history, and leadership backgrounds
Structures findings into a standardized memo format
Generates follow-on research questions based on initial findings

What you do:

Define the scope and priorities with the client
Add industry context and judgment that AI cannot supply
Verify claims that require extra confirmation
Deliver and discuss findings with decision-makers

Tools: Claude Opus 4.6, Perplexity for research, Make.com for intake, Notion for delivery.

Revenue: $1,000 to $5,000 per engagement. Retainer clients at $5,000 to $15,000 per month.

16. AI Bookkeeping Synthesis Service

What it is: A financial intelligence layer for small businesses that have a bookkeeper but no one who can interpret the numbers. You take structured financial data and produce clear, actionable insights.

What the AI does:

Ingests categorized financial data from QuickBooks or Xero
Identifies trends, anomalies, and cost drivers
Generates plain-language summaries for non-financial founders
Flags questions the founder should be asking their accountant

What you do:

Review AI output for accuracy
Add context and interpretation for each specific business
Deliver monthly briefs and answer follow-on questions
Build long-term financial context over time

Tools: Claude Opus 4.6 for financial reasoning, QuickBooks or Xero API for data access, Notion for delivery.

Revenue: $500 to $2,000 per month per client. High margin. Low delivery cost once the workflow is set.

How to Pick Your One-Man Business

Do not pick the idea with the highest revenue ceiling. Pick the one where your existing skills cover what AI cannot do yet.

AI handles volume, research, drafting, monitoring, and generation. It does not handle client judgment, creative direction, relationship trust, or accountability.

Pick the idea where your human contribution is the actual value. The one where clients would not trust the output without a person behind it.

If you have...	Start with...
Deep expertise in a subject	AI Research Boutique or Due Diligence Service
Sales or outbound experience	AI Outbound Sales Operation
Content or marketing background	AI Content Factory or AI Newsletter Business
Finance or accounting background	AI Financial Reporting or Bookkeeping Synthesis
Technical skills and industry knowledge	Vertical AI SaaS Product
Recruiting or HR experience	AI Recruiting Pipeline
Creative direction and taste	AI Social Media Ghost-Operator

If you are not sure where to start, the do-nothing.ai scoreboard shows real people running AI-powered one-person businesses right now. The model works. The question is which version of it fits what you know.

Before You Launch: The Foundation Every Idea Shares

Every business on this list runs on the same underlying stack:

A core LLM. Claude Opus 4.6 for complex reasoning. Claude Haiku 4.5 for volume tasks. Match the model to the task and your cost structure stays manageable.
An automation layer. Make.com or n8n. This is the plumbing that connects your tools and makes the business run without you touching every step.
A way to store client context. Notion for early stage. A proper database as you scale. Agents produce better output when they have structured context about who they are working for.

For the full stack setup, start with the 2026 Solopreneur AI Stack. For how to actually charge for any of these, read How to Price an AI-Powered Service before you quote a single client.

Related Guides

10 AI Agent Business Ideas You Can Start Solo in 2026

neo one — Fri, 20 Mar 2026 12:59:22 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

10 AI Agent Business Ideas You Can Start Solo in 2026

AI helpers aren't just a feature you add to an existing business. They can be the whole business.

An AI agent is software that thinks, takes actions, and completes multi-step tasks without someone holding its hand at every step. When you build a business around AI agents, you're selling the results of that work at scale. One person can operate what looks like a team.

These are not hypothetical ideas. These are real business models with real demand, real pricing, and clear paths to launch. For each one: what the business does, who the customer is, what tools you need, realistic earnings, and how long it takes to start.

Before / After — Plain English

Before: "You deploy agents that do deep research on any topic: competitive landscapes, market sizing, due diligence on companies, industry primers."

After: "You set up AI to do in-depth research on any topic and deliver a structured report. You review and send it."

1. AI Research Service

What it is: You set up AI to do deep research on any topic — competitor landscapes, market sizing, company due diligence, industry overviews. The AI searches, organizes, and produces a structured report. You review, polish, and deliver.

Who it's for: Private equity analysts, startup founders, consultants, journalists — anyone who bills for research they currently do by hand.

Tools needed: Claude (for reasoning), Perplexity or Exa (for web search), Make.com (to wire intake forms to delivery), Notion (for report templates).

Earnings: $300–2,500 per report depending on scope. Manual research boutiques charge $5,000–25,000. You sit below that and well above freelancer rates.

Time to launch: 2 to 4 weeks. Build one AI workflow. Test on five sample briefs. Refine the output. Start with three beta clients at a discount.

2. Automated Lead Generation Pipeline

What it is: You build and run an outbound sales system for B2B companies. AI finds target companies, gets contact info, and writes personalized outreach messages. These go straight into the client's inbox or sales system.

Who it's for: B2B companies with 5 to 100 employees that know they need outbound sales but don't have a dedicated salesperson.

Tools needed: Clay (for finding leads), Claude (for personalizing messages), Make.com (for automation), Instantly or Apollo (for sending emails).

Earnings: $2,000–8,000 per month per client. Clients compare this to hiring a full-time salesperson at $50,000–70,000 per year. You win on price and speed.

Time to launch: 3 to 6 weeks. The pipeline setup takes time. Once built, adding new clients takes little extra effort.

3. AI Content Operations for Software Companies

What it is: You run a content machine for software companies that need regular blog posts, guides, and landing pages but can't justify a full content team. AI handles keyword research, first drafts, and internal linking. You handle editing, strategy, and client communication.

Who it's for: Software companies earning $1M–10M per year. They have budget, they know content matters, and they don't want to hire a content manager.

Tools needed: Ahrefs (keyword research), Claude (drafting), Notion (tracking), Make.com (workflow), Vercel or WordPress (publishing).

Earnings: $3,000–12,000 per month per client. Content agencies charge $5,000–20,000. You deliver faster with less overhead.

Time to launch: 4 to 6 weeks. Build the workflow once. Customize intake per client.

4. AI-Powered Customer Support

What it is: You build and run a first-response support system for small businesses. AI handles common questions — password resets, order status, FAQs, basic complaints. Anything it can't handle gets sent to a human.

Who it's for: Online stores, software startups, and service businesses with high support volume and a small team.

Tools needed: Claude (the fast, cheap version for responses), Intercom or Zendesk (support platform), Make.com (for routing), Supabase (to store client's knowledge base).

Earnings: $1,500–6,000 per month per client. Clients compare this to hiring a part-time support person. You handle more volume at a fraction of the cost.

Time to launch: 3 to 5 weeks. Loading the client's knowledge base and testing takes the most time.

5. AI Financial Reporting Service

What it is: You build AI that takes raw financial data — payment records, bookkeeping exports, bank statements — and produces clear reports: monthly income summaries, cash flow analysis, runway projections.

Who it's for: Startups and small businesses that have a bookkeeper but no financial director. They have the data. They don't have the analysis.

Tools needed: Claude (for financial reasoning), Python or Make.com (for pulling in data), Notion or Google Slides (for report delivery).

Earnings: $1,000–5,000 per month per client. A part-time financial advisor charges $2,000–10,000. You sit at the lower end with faster output.

Time to launch: 4 to 8 weeks. Normalizing data from different sources is the hard part. Once solved for one client type, it repeats.

6. AI Recruiting Pipeline

What it is: You build an AI-powered hiring workflow for companies that hire frequently. AI screens incoming resumes, writes outreach to passive candidates, and sends screening questions. Humans only see candidates who make it past the filter.

Who it's for: Hiring managers doing five or more hires per quarter who find the screening process brutal.

Tools needed: Claude (for resume review and outreach drafting), Make.com (for connecting to their hiring system), Apollo (for finding candidates).

Earnings: $3,000–10,000 per month per client. Traditional recruiters charge 15–25% of first-year salary. You charge a flat monthly fee and save them significantly on every hire.

Time to launch: 4 to 6 weeks. Customizing per client's hiring criteria takes time. The core workflow reuses across clients.

7. AI Monitoring and Alerts Service

What it is: You build AI that watches specific signals for clients — competitor price changes, industry news, regulatory updates, social media mentions, or job postings at target accounts. When something relevant surfaces, the AI summarizes it and sends an alert.

Who it's for: Sales teams tracking competitors, compliance teams watching regulatory news, investors following their portfolio companies, PR teams monitoring brand mentions.

Tools needed: Exa or Perplexity (for web monitoring), Claude (for summarizing), Make.com or n8n (for scheduling and delivery), Slack or email (for alerts).

Earnings: $500–3,000 per month per client. Low delivery cost makes this high margin once set up.

Time to launch: 2 to 3 weeks. This is the fastest to launch on this list. Build an MVP for one type of client and expand from there.

8. AI Contract and Document Review

What it is: You offer a service that reviews standard business documents using AI — NDAs, vendor contracts, employment agreements, terms of service. The AI flags unusual clauses, missing provisions, and high-risk language. You help non-lawyers know what questions to ask their lawyer.

Important: This is a first-pass review tool, not legal advice. Be clear about that with clients.

Who it's for: Founders and small business owners who see contracts constantly and can't afford to run every one by an attorney.

Tools needed: Claude (the most capable version, for nuanced document reasoning), a secure file upload system, Notion or PDF for report delivery.

Earnings: $200–800 per document, or $1,000–4,000 per month for clients with regular document flow.

Time to launch: 3 to 5 weeks. Most setup is crafting good instructions and building a library of what to look for in each document type.

9. AI Social Media Content Service

What it is: You run social content for clients who need consistent posts on LinkedIn, X (Twitter), or Instagram. AI generates ideas based on the client's expertise, drafts posts in their voice, and queues them for your approval. You handle the strategy and final review.

Who it's for: Founders, executives, and B2B companies who understand their audience expects regular content but don't have time to produce it.

Tools needed: Claude (for drafting in brand voice), Buffer or Hypefury (for scheduling), Make.com (for approval workflow).

Earnings: $1,500–5,000 per month per client. Social media managers charge $2,000–6,000. You match quality at lower cost because AI handles production.

Time to launch: 2 to 4 weeks. Matching the client's voice takes the most time. Once dialed in, monthly delivery is fast.

10. Vertical AI Tool for One Industry

What it is: Instead of a service, you build a software product for one specific industry — real estate agents, dentists, law firms, accountants, or contractors. The AI handles the repetitive work specific to that industry. You sell it as a subscription.

Who it's for: Professionals in the target industry who aren't technical and just want the outcome, not the setup.

Tools needed: Claude (the cheap, fast version for volume), Vercel (for hosting), Supabase (for user accounts and data), Stripe (for subscriptions).

Earnings: $99–499 per month per customer. At 100 customers, that's $10,000–50,000 in monthly recurring revenue — from a product that runs without you. Highest ceiling on this list.

Time to launch: 8 to 16 weeks. This is the most complex starting point. Don't begin here unless you know the target industry well.

How to Pick the Right Idea

Don't pick the idea with the highest earning ceiling. Pick the one that matches what you already know.

If you have...	Start with...
Deep expertise in a subject	AI Research Service or Document Review
Sales or outbound experience	Lead Generation Pipeline
Content or marketing background	Content Operations or Social Media
Finance or accounting background	Financial Reporting Service
Technical skills and industry knowledge	Vertical AI Tool (the product model)
Recruiting or HR experience	AI Recruiting Pipeline

The fastest path to income is not the idea with the most potential. It's the idea where your existing skills cover the 20% of work AI still can't do reliably. AI handles volume and consistency. You handle judgment.

What Every One of These Businesses Needs

Every option on this list runs on a shared foundation:

A core AI model. Claude (the smart version) for complex reasoning. Claude (the fast version) for high-volume, simpler work. Matching the right version to the task controls your costs.
An automation connector. Make.com links your tools together without code. Once you know it, most workflows come together in days.
A way to manage client information. Notion works early. As you grow, you need a real system for storing client details and preferences so AI produces consistent results.
A pricing strategy. Most first-time AI service founders underprice by 30–60%. Read How to Price an AI-Powered Service before quoting a single client.

For the full launch playbook, see How to Start an AI Business. For structuring deals so you build real profit — not just revenue — see AI Business Monetization.

Related Guides

How to Make Passive Income with AI (What Actually Works in 2026)

neo one — Fri, 20 Mar 2026 12:54:03 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

How to Make Passive Income with AI (What Actually Works in 2026)

Passive income is not passive. Someone lied to you. What people actually mean is income that does not require proportional time. You do not get paid for every hour you work. You build a system, and the system earns.

AI makes that possible at a much smaller scale than it used to require. You do not need a team, a warehouse, or a venture round. You need a system that can run without you touching it every day.

This guide covers the models that actually work. Not "use ChatGPT to write an ebook." The real ones.

What Makes AI Income "Passive"

A traditional passive income model requires upfront capital or content. You buy a rental property. You write a book. You record a course. Then you wait.

AI changes the calculus. Agents can produce content at scale, handle customer interactions, generate personalized outreach, run support, and manage distribution. The upfront work is building the system. After that, the system runs.

The result is not completely passive. Edge cases still need human judgment. Strategy still requires thinking. Systems break and need maintenance.

But the ratio of time-in to money-out is genuinely different from a traditional service business or a job.

The Models That Work

1. AI-Powered Content Sites

You build a content site on a specific topic. AI agents produce the content. The site earns from display ads, affiliate links, or sponsored placements.

The honest version: this worked better before 2024. Search traffic to generic content sites has fallen as AI summaries eat into click-through rates. What still works: niche sites on specific, high-intent topics where the content is genuinely more useful than what AI generates in a summary box.

The right approach in 2026: focus on comparison content, buyer's guides, and data-heavy resources that require original research or aggregation. Agents can draft the structure. Your editorial layer makes it useful enough to rank and convert.

Realistic income range: $500 to $5,000 per month per site at steady state. Requires 6 to 18 months to reach meaningful traffic. If that timeline makes you flinch, skip to the service business section. It is faster and less existential.

2. AI-Driven Info Products and Micro-SaaS

You build a tool or resource that solves a specific problem. AI is the engine. Customers pay for access.

Examples that work:

A niche AI assistant that does one specific task for one type of user (e.g., proposal generator for freelance designers)
A research digest that aggregates and summarizes news for a specific industry, delivered weekly by email
A template library with an AI assistant that customizes templates on demand
A lightweight SaaS tool that automates one painful data task for a specific job title

The key: narrow market, real problem, AI makes the solution cheaper to deliver than a human-powered alternative.

Realistic income range: $300 to $10,000 per month depending on pricing and market size. Requires actual distribution to an audience that wants it.

3. Automated Service Businesses

You sell a service. AI does most of the delivery. You do quality control and client strategy.

This is the most practical passive income model for a solo operator in 2026. "Passive" is relative here because you are still running a business, but the marginal effort per client is very low.

Examples:

SEO content production: agents draft, you edit, clients get volume
Automated competitive intelligence reports: agents pull data and summarize, you package and sell subscriptions
AI-assisted outbound lead generation: agents handle prospecting and personalization, you handle strategy

This model can scale to $10,000 to $50,000 per month for a solo operator with strong systems and good clients. The ceiling is real. But so is the floor: you can have clients and revenue within weeks, not months.

4. Licensing Agent Workflows and Automation Templates

You build a workflow that works. You sell access to it.

This is underrated. Most businesses need automation but do not know how to build it. A well-built Make.com scenario or n8n workflow that solves a specific problem can be sold as a template, a done-for-you setup, or a license.

Examples:

A complete customer onboarding automation for a specific industry
A lead qualification and routing system for a specific type of sales team
An automated social media content pipeline for a specific audience

Pricing varies widely. Templates sell for $50 to $300 one-time. Done-for-you setups: $500 to $3,000. Licensing with support: monthly recurring.

5. API-Powered Products

You build something on top of the AI APIs that is genuinely useful. Not a prompt wrapper. An actual product with a real workflow underneath it.

This requires more technical skill than the other models. But the ceiling is higher. API-powered products can reach large audiences because the cost to serve an additional user is low.

Examples: summarization tools, document analysis tools, research assistants for specific verticals.

The trap: building a product that is one API update away from being a native feature of the underlying model. Build for specific workflows, not general capabilities.

What Does Not Work

AI ebooks sold on Gumroad. The market is saturated. Buyers know. Move on.

Generic content farms with AI-generated articles. Google has adjusted. Thin AI content without real editorial value does not rank. This does not mean AI content cannot rank. It means low-effort AI content cannot.

Prompt marketplaces. The market peaked. Prompts are commoditized.

Dropshipping with AI product descriptions. The margins do not justify the effort unless you have a very specific edge in sourcing or a real brand.

Any model that claims you do nothing at all. There is always a system to maintain, customers to handle, and strategy to set. The passive part is the execution layer. The rest is still work.

The Real Upfront Investment

Passive income is not passive in setup. The honest list of what it takes:

Time to build the system. Two weeks to three months depending on complexity. This is full-effort work. Building automations, testing workflows, refining outputs.

Tool costs before revenue. API costs, Make.com or n8n subscription, hosting, possibly a domain and simple website. Budget $100 to $400 per month for a minimal stack.

Distribution before a product earns. You need an audience or a marketing channel before a product generates passive income. This is often the missing piece. The automation can run perfectly and still earn nothing if no one finds it.

Patience. The first three months of most AI income models are not passive. They are active and often unprofitable. The passive part kicks in after the system is working and distribution is consistent.

The Smartest Path in 2026

Start with an automated service business. Get real revenue quickly. Use that revenue to fund building a product. The service builds your understanding of what customers actually need. The product captures leverage.

Many of the most successful solo operators running AI-powered income follow this exact sequence. They start by doing work (efficiently, with AI), then they productize what works.

The "skip straight to passive" fantasy costs time. The service-first path builds a business. Build the income first. Let the passivity arrive later.

Related Guides

AI Agents for Small Business: A Practical Playbook

neo one — Fri, 20 Mar 2026 12:53:56 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

AI Agents for Small Business: A Practical Playbook

Running a small business is not a problem of ideas. It is a problem of bandwidth. You have five people doing the work of fifteen, and half of that work is repetitive enough that a well-configured AI agent could handle it while you sleep.

This guide is for small business owners with one to five employees. Not enterprise IT rollouts. Not venture-backed startups with a full engineering team. Real businesses, limited headcount, real stakes.

Here is what AI agents actually do in small businesses today, and how to deploy them without wasting a month on setup.

What AI Agents Are (and Are Not)

An AI agent is software that takes a goal, uses tools to work toward it, and produces an output without you babysitting every step.

Agents in 2026 are reliable for:

Drafting and editing: first-pass emails, proposals, social posts, responses to customer inquiries
Research and summarization: competitor monitoring, industry news, product research
Classification and routing: sorting inbound messages by type, scoring leads, tagging support tickets
Data entry and transformation: moving information between systems, reformatting, deduplicating
Scheduling and follow-up: reminders, follow-up sequences, booking confirmations

Agents are not reliable for:

Decisions that require judgment about a specific customer's history or feelings
Anything where being wrong creates a costly or public problem
Creative work that needs genuine taste, not pattern-matching on what has worked before

The small business advantage is this: most of your repeatable work fits squarely in the first list.

The Four Functions to Automate First

1. Marketing

The problem: You know you need to post consistently, send the newsletter, and stay top of mind. But marketing is always the first thing to slip when you are busy running the actual business.

What an agent handles:

Drafts weekly social content from a topic list you set once a month
Turns a blog post into five social posts and a newsletter section
Monitors competitors and surfaces what they are doing (without you checking their sites manually)
Schedules and queues content through your existing tools

What you still own: the strategy, the final review before anything goes live, and any content that touches sensitive topics or big announcements.

Simple setup: Claude + Make.com + Buffer. You approve. The agent publishes.

2. Customer Support

The problem: Customers email at 11pm. They want answers. You cannot be on call 24 hours, and hiring a support rep for 40 questions a week does not pencil out.

What an agent handles:

Answers common questions from a knowledge base you configure once
Triages inbound messages by urgency and type
Drafts responses to complex questions for your review before sending
Escalates anything that looks like a complaint, refund request, or legal issue

What you still own: the escalations, the edge cases, the relationship-critical interactions.

Simple setup: Claude + your email provider or Intercom + a knowledge base doc. The agent drafts. You send.

3. Operations

The problem: Ops is the tax on running a business. Scheduling, reminders, status updates, internal coordination. Everyone on the team spends time on it. Most of it is automatable.

What an agent handles:

Weekly status summaries pulled from your project management tool
Meeting notes drafted from a transcript
Invoice reminders sent on a schedule
New client onboarding emails triggered by a form submission

What you still own: the judgment calls, the interpersonal stuff, anything that requires reading a room.

Simple setup: Make.com + Claude + your existing tools (Notion, Airtable, Google Workspace).

4. Bookkeeping Prep

The problem: Bookkeeping is not hard. It is tedious. Categorizing expenses, reconciling receipts, generating monthly summaries. Your accountant charges by the hour. You are doing the easy parts yourself.

What an agent handles:

Pulls and categorizes transactions from your bank export
Flags anomalies or uncategorized items for your review
Generates a plain-language monthly summary you can hand to your accountant
Matches receipts to transactions when you forward them by email

What you still own: the final sign-off, tax strategy, and anything that requires a CPA.

Simple setup: Claude + your accounting software's export + a simple review workflow. This alone saves most small businesses 2 to 4 hours a month.

How to Start Without Building a Custom System

You do not need to write code. You do not need a developer. You need three things:

1. A trigger (what starts the agent): a new email, a form submission, a scheduled time, or you typing a prompt.

2. An action (what the agent does): calls Claude with a clear instruction and relevant context.

3. An output (where the result goes): an email draft, a Notion page, a Slack message, a spreadsheet row.

The simplest stack for a small business:

Trigger (Make.com webhook or cron)
  -> Claude (Sonnet or Haiku depending on task complexity)
  -> Output (email, doc, spreadsheet, notification)

Make.com has a library of prebuilt connectors. You configure the logic visually, no code required. Claude handles the language tasks. You review and approve what matters.

Start with one task. Pick the most repetitive thing your team does that has clear inputs and outputs. Build that first. Run it for two weeks. Then add the next one.

What to Expect in the First 30 Days

Week 1: Setup takes longer than you expect. The agent's first drafts are 80% of the way there, not 100%.

Week 2: You spend time editing. This is normal. The agent is learning your patterns through the instructions you refine.

Week 3: You start trusting the drafts more. Your edit time drops. You start thinking about what to automate next.

Week 4: You have a running system. You are not thinking about the agent. It is just part of how the business works.

The first month feels like investment. The second month pays off.

The Mistakes Small Businesses Make

Automating without a human review step. The most common failure mode. An agent sends something slightly wrong to a customer. You did not catch it because you removed the review step to save time. Keep the review step.

Building too much at once. Two-week setup, zero usage. Start with one task. Ship something that works in days, not weeks.

Using the wrong tool for the task. Claude Opus for categorizing 200 invoices is expensive and unnecessary. Claude Haiku handles classification tasks at a fraction of the cost. Match the model to the task complexity.

Not writing good instructions. The agent's output quality is a direct function of how well you described the task. Spend more time on your prompt than on your workflow configuration.

Cost Reality

A small business running three agent workflows (marketing drafts, support triage, ops summaries) typically spends $20 to $80 per month on API costs depending on volume.

Compare that to:

A part-time marketing coordinator: $1,500 per month
A customer support contractor: $800 per month
8 hours of your own time per week at your effective hourly rate

The math works. The barrier is setup, not ongoing cost.

Use the do-nothing.ai calculator to estimate your specific time savings and ROI before you build anything.

Getting Started This Week

List every task your team does more than twice a week
Circle the ones with the clearest inputs and outputs
Pick the one that takes the most total time across the team
Build a single-step agent for it using Make.com and Claude
Run it for two weeks before expanding

You are not automating your whole business on day one. You are removing one constraint. Then another. Then another.

That is how small businesses end up running like large ones, with fewer people and far less chaos.

Related Guides

How Solopreneurs Use AI Agents to Scale Without Hiring

neo one — Fri, 20 Mar 2026 12:47:56 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

How Solopreneurs Use AI Agents to Scale Without Hiring

You're one person running a business. Your calendar says you need three of you by Thursday.

AI helpers don't do everything. But they handle the repetitive, time-consuming work — so you can focus on the 20% that actually requires your judgment.

This guide is for people running real businesses alone. Not demos. Not toy projects. Revenue-generating businesses where time is the real bottleneck.

Before / After — Plain English

Before: "Agents are reliable for research and synthesis, draft generation, classification and routing, monitoring and alerts, and repetitive transformation."

After: "AI helpers are great at: research, writing first drafts, sorting incoming requests, and doing the same task over and over."

What AI Agents Do Well (and What They Don't)

AI helpers in 2026 are reliable for:

Research and summaries — pulling information from multiple sources and organizing it
First drafts — emails, proposals, articles, code
Sorting and routing — categorizing incoming requests, tagging data, deciding what goes where
Watching and alerting — noticing when things change and telling you when something needs attention
Repetitive tasks — taking one thing and converting it to another (reformatting, translating, restructuring)

AI helpers still struggle with:

Making new decisions with incomplete information
Tasks that require knowing a specific customer's history or personality
Anything where a mistake is costly and hard to detect
Creative work that needs genuine taste, not just pattern-matching

If your task requires "you had to be there" context, the AI wasn't there. Handle it yourself.

Four AI Patterns That Work for Solo Business Owners

1. The Research Helper

Problem: You need competitive information, market data, or background before writing or deciding anything. This used to take hours.

How it works:

You send a research question
AI browses sources, pulls key data, removes noise
You get a structured summary with sources
You review, add context, act

Tools: Claude + web browsing + structured output

2. The Content Pipeline

Problem: You need to publish content regularly (blog posts, newsletters, social posts) without it becoming a full-time job.

How it works:

AI gets a list of topics
AI drafts each piece
You review and edit (15–30 minutes instead of 2 hours)
AI formats and schedules the published version

Important: Never skip the review step for anything customer-facing. The AI's job is to make your editing faster, not to replace it.

3. The Inbox Triage Helper

Problem: Email, support tickets, and new leads all need first-response handling. You can't be monitoring all day.

How it works:

A new message comes in
AI reads it and figures out what it is (support, sales inquiry, partnership, spam)
For support: AI drafts a response for your review
For new leads: AI finds more info and adds them to your list
For spam: AI archives it automatically

Tools: Make.com + Claude + Airtable

4. The Monitoring Helper

Problem: You need to know when things change — competitor prices, news about your industry, key numbers shifting — without checking dashboards all day.

How it works:

AI runs on a schedule (hourly, daily, weekly)
Pulls data from sources you chose
Compares to what it saw before
Only sends you a message when something actually matters

Tools: Vercel Cron + Claude + Resend

How to Structure Your AI Setup

Start Small

Don't build a complex multi-AI system on day one. Start with one helper that solves one real problem you have right now.

The most common mistake: spending two weeks building an AI pipeline and then not using it because it doesn't fit your actual workflow.

Use a Simple Structure

For most solo businesses:

Trigger (a schedule / a webhook / you)
  → Manager AI (Claude Sonnet: routes and decides)
  → Worker AIs (Claude Haiku: execute specific tasks)
  → Output (stored, emailed, or logged)

You only need more complexity when this breaks down.

Note: Claude Haiku is the small, fast, cheap version of Claude. Claude Sonnet is the mid-range version that thinks more carefully.

Keep Humans in the Loop For These

Any outbound message to a real customer or lead
Financial decisions over a threshold you set
Anything that creates a public commitment (posts, contracts, proposals)
Decisions involving people (hiring, partnerships, conflicts)

Automate the preparation. Keep the send button for yourself.

Measuring Whether It's Worth It

Track this simply:

Time saved per run — how long did this task take you before?
Cost per run — check your Claude usage
Runs per month — how often does it fire?

If time saved × your hourly value > cost × runs per month, the AI pays for itself. Most do, quickly.

Getting Started This Week

List the three tasks that eat the most of your time each week
Pick the one that is most repetitive and has the clearest inputs and outputs
Build a simple single-helper system for that task
Run it for two weeks and measure time saved
Then add the next one

You are not building a robot army. You are building a reliable assistant that shows up every day, never calls in sick, and doesn't need a pep talk Monday morning.

Start small. Let the results sell you on adding more.

Related Guides

How to Make Money with AI in 2026 (What Actually Works)

neo one — Fri, 20 Mar 2026 12:47:50 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

How to Make Money with AI in 2026 (What Actually Works)

Scroll through any social feed and someone is promising you can "make $10k a month with AI." They're usually selling a course.

This guide is free. And honest.

There are four real paths to income using AI in 2026. Each one works. None of them are easy. And every single one requires something AI cannot do for you: finding people who will pay.

For the full list of AI income options, see Best AI Side Hustles in 2026. This guide focuses on the four highest-leverage paths and how to choose between them.

Before / After — Plain English

Before: "AI dramatically lowers execution costs, but it does not create moats by itself."

After: "AI makes things cheaper and faster. But it doesn't find you customers or make your offer worth buying."

What It Actually Means to Use AI to Make Money

Using AI to make money is not about the tools. It's about what the tools let you do that you couldn't do before.

Before 2024, one person could produce maybe 5 to 10 quality articles per month. Now they can produce 30 to 50 at the same quality. Before 2025, building a software product required a technical partner or a big budget. Now a non-developer can ship a working product in a week.

AI makes things cheaper and faster. But AI doesn't find you customers. It doesn't price your services correctly. It doesn't build the relationships that turn a first project into regular income.

Most people fail to make money with AI not because they used the wrong model. They fail because they treat AI as the business instead of as the engine inside a business.

Get that mental model right first. Then the tool choices become obvious.

The Four Real Paths to Income Using AI

In 2026, there are four categories where people are actually generating real income:

Freelancing and services — use AI to deliver faster and better, charge for the output
Content and SEO — use AI to produce at scale, earn from traffic or from clients who need content
Consulting and implementation — charge for knowing how to deploy AI for other businesses
AI agent business — build a system where AI does the work; you collect the revenue

Each path has a different ceiling, a different time to first dollar, and requires different skills. The right path depends on what you already have: an audience, specific skills, technical ability, or money to invest.

Path 1: Freelancing — Fastest Way to Start

Freelancing with AI means using AI tools to do more work faster — not replacing your judgment with the machine's.

Clients in 2026 are not paying for access to Claude or ChatGPT. They're paying for your ability to produce good work fast and take responsibility for quality. Anyone can prompt an AI. Not everyone delivers work that meets a professional standard.

What this looks like in practice:

A freelance writer who used to deliver 2 articles per week now delivers 6 to 8. They use AI to synthesize research, draft structure, and generate first passes. They spend their time editing, checking accuracy, and matching the client's voice.

A developer uses Claude Code to write boilerplate and generate tests. They take on projects that used to require a small team.

Earnings: $3,000–15,000 per month for a focused solo freelancer. AI moves the ceiling up significantly but doesn't eliminate it.

Time to first dollar: 1 to 3 weeks, assuming you have a skill and can reach potential clients.

The trap: Using AI to compete on price instead of quality. If you offer AI-assisted writing at $25 per article, you'll find clients — but you'll be stuck at a rate that makes real income impossible. Position AI as a quality and speed advantage, not a price justification.

Path 2: Content and SEO — Using AI to Make Money at Scale

Content is one of the most direct paths to income with AI — both as a service and as something you own.

Content as a service: Produce content for other businesses. AI handles drafting and research. You handle editorial quality and client relationships. A solo content operator with five clients can produce 60 to 80 pieces per month. At $150–300 per piece, that's $9,000–24,000 per month.

Content as an asset: Build your own website or newsletter where AI-assisted content earns through ads, referral fees, or product sales. This takes longer but compounds. Content you publish today can earn for years.

In 2026, the content market has split in two: cheap, undifferentiated volume (low value) and high-quality structured content that ranks, earns trust, and gets cited (high value). The income is in the second category.

Earnings: $5,000–20,000 per month for a service operator. Owned content sites vary widely — a well-executed niche site can earn $3,000–30,000 per month by year two or three.

Time to first dollar: 1 to 2 weeks for a service (once you have a client). 3 to 6 months for an owned site (once you have meaningful traffic).

Path 3: Consulting — Charge for Knowing What Works

Most businesses in 2026 know they need to use AI better. Very few have figured out how. The gap between "we should be using AI" and "here is our working AI workflow" is where consultants earn.

Advisory work: Workshops and strategy sessions for business teams. Day rates for workshops run $3,000–15,000 depending on your positioning.

Implementation work: You build the actual workflows. You connect the tools. You hand over something that runs. This is more valuable and more durable. Monthly retainers of $3,000–10,000 for ongoing implementation support are common.

What makes this work: Credibility comes from having actually built things, not from knowing the terminology. The consultants earning real money have built agent workflows, automated content pipelines, or AI-assisted sales processes — and can show the results.

Warning: Basic AI training ("how to use ChatGPT") is already a commodity. Implementation consulting is not.

Earnings: $5,000–25,000 per month for an established implementation consultant.

Time to first dollar: 1 to 3 weeks if you have a network to sell into.

Path 4: The AI Agent Business — Highest Ceiling

This is the path most guides skip because it takes longer to explain and doesn't produce income in week one.

An AI agent business is one where AI helpers handle most of the operating work — writing content, running outreach, handling support, doing research, managing workflows. You design the system. The AI runs it. Revenue comes from the product or service the system produces, not from your hourly output.

The do-nothing.ai model is the clearest version: a one-person operation that runs at the output of a small team because AI is doing the repetitive, parallelizable work. The founder makes decisions, sets direction, and handles relationships.

Why this has the highest ceiling: Service businesses cap at what one person can manage and charge. AI agent businesses can serve more customers without adding more hours. The marginal cost of serving one more customer approaches zero.

Some solo operators running this model earn $500,000+ per year without a single employee.

What you need:

A product or service with repeatable, definable workflows
Enough technical comfort to set up and connect AI tools
A clear enough offer to get customers before the system is fully built
Patience — this model takes 2 to 8 weeks before you see the first dollar

Earnings: Highly variable. Early months may be $0–3,000/month while the system matures. By month 6–12, functional AI businesses in good markets reach $5,000–30,000/month.

Time to first dollar: 2 to 8 weeks depending on complexity and how fast you find customers.

For a full walkthrough, see How to Start an AI Business with No Team (2026).

How to Choose the Right Path

The right path depends on what you have right now:

If you have an existing skill and some network: Start with freelancing or consulting. First income in 1 to 3 weeks. Use that revenue and those conversations to decide whether to build a product later.

If you have writing or marketing skills but no clients: Content as a service is accessible. One client with consistent volume makes the economics work.

If you have technical ability and can build workflows: The AI agent business model is worth the extra ramp. The ceiling is higher and the business compounds.

If you have an existing audience: Content assets and digital products are faster for you than for someone starting from zero.

Most people should start with a service (immediate income) and build toward an AI-run product (compound value). The service pays your bills and teaches you what customers actually need. The product is where leverage compounds.

The Distribution Problem

Every path on this list requires customers. AI speeds up the work. It doesn't find you buyers.

This is what most guides skip. They walk you through tools and workflows and leave out the harder problem: how do you get in front of people who will pay?

The fastest ways to find customers in 2026:

Existing network: The fastest path to first income is always selling to people who already know you. Announce what you're doing. Be direct about who you help.
Cold outreach: Research leads with AI, write personalized emails with AI. This is a skill that takes 30 to 60 days to get working well.
Build in public: Writing about what you're building while you build it generates inbound interest before you need it.
Referrals: The most reliable ongoing source for service businesses. One happy client who tells three others is worth more than any ad.

Don't build a complex product in private for months before anyone sees it. Distribute as you build. The feedback makes everything better faster.

What to Do First

If you're starting today with no customers:

Pick one path based on what you already have
Name five specific people or companies you could sell to
Reach out to all five before you build anything
Let the first interested response tell you what to build

AI makes iteration cheap. Use that. Start rough, sell early, improve fast. Nobody hands you revenue for having a good setup. Go sell something.

Related Guides

Best Vibe Coding Tools in 2026

neo one — Fri, 20 Mar 2026 09:47:44 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

Best Vibe Coding Tools in 2026

Vibe coding is the fastest-growing practice in the solopreneur and indie developer world. The core idea: you describe what you want in plain language, the AI builds it, and you iterate on the result instead of writing every line yourself.

The term was coined by Andrej Karpathy in early 2025 and it spread fast because it described something real. A new mode of building software where your job shifts from typing code to directing AI, reviewing output, and making judgment calls.

This guide covers the six tools solopreneurs are actually using for vibe coding in 2026, what each one is best at, and how to choose.

What Is Vibe Coding?

Vibe coding is a workflow where AI writes the majority of the code. You provide intent. The AI provides implementation. You review, refine, and redirect.

This is different from AI-assisted coding, where you still write most of the code and AI helps with completions or suggestions. Vibe coding flips the ratio. You spend more time reading and steering than typing.

For solopreneurs, this matters because:

You can build faster without a team
You can work outside your primary stack (frontend dev building a backend, for example)
The quality bar for "good enough to ship" is reachable without senior engineering expertise
Iteration cycles shrink from days to hours

The tradeoff is that vibe coding requires you to develop a new skill: knowing what to ask for, how to review AI output critically, and when to override the AI's choices.

Who Vibe Coding Is For

Vibe coding is a strong fit if you are:

A solopreneur building a web app, SaaS, or internal tool
A non-technical founder who wants to prototype without hiring
A developer who wants to move 3-5x faster on routine features
An operator who wants to automate something but does not want to pay for a developer

It is not a good fit if you are:

Building systems where correctness is critical (finance, medical, infrastructure)
Working on codebases where the AI cannot understand enough context to be useful
Expecting zero-review output you can deploy without reading

With that framing, here are the six best vibe coding tools available in 2026.

Cursor

What It Is

Cursor is an AI-first code editor built on VS Code. It keeps everything from VS Code and adds a deeply integrated AI layer on top. As of 2026, it supports Claude, GPT-4o, and other frontier models.

For vibe coding, Cursor's Agent mode (called Composer) is the primary feature. You describe a task, the agent reads your codebase, writes and edits files across multiple directories, runs terminal commands, and iterates until it's done.

Pricing

Free: 2,000 completions/month, 50 slow premium requests
Pro ($20/month): 500 fast premium requests, unlimited slow requests
Business ($40/user/month): centralized billing, admin controls

Best For

Developers who already use VS Code and want zero migration friction
Day-to-day coding with AI as a persistent co-pilot
Feature development where you want to stay in the editor while the AI works
Teams that need centralized billing

Strengths

Tab autocomplete is the best in class for in-file suggestions
Agent mode handles complex, multi-file tasks with strong accuracy
Supports @-mentioning specific files, folders, docs, and URLs as context
Large and active user community with well-documented workflows

Weaknesses

Premium request limits can be hit quickly during intense sessions
$20/month Pro tier is the minimum for serious use
Agent mode quality drops on very large or deeply interconnected codebases

Bottom Line

Cursor is the best vibe coding tool for developers who want to stay in an IDE environment. The combination of autocomplete and agent mode gives you two distinct leverage points: AI assistance while you type, plus full task delegation when you want it.

Windsurf

What It Is

Windsurf is Codeium's AI IDE, also built on VS Code. Its differentiator is Cascade, an agentic flow system built for codebase-wide changes. Rather than assisting you file by file, Cascade reasons about your entire architecture before taking action.

Pricing

Free: unlimited Codeium completions, 5 Cascade Flow credits/month
Pro ($15/month): 500 Flow credits/month, access to frontier models
Pro Ultimate ($35/month): 1,500 Flow credits, higher caps

Best For

Large-scope changes: migrations, refactors, adding a feature across many files
Cost-sensitive solopreneurs who want strong AI capability at a lower price
Developers who want the agent to plan and execute a full task with minimal interruption

Strengths

Cascade Flow handles multi-file, architecture-aware changes better than any other IDE tool
Lower base price than Cursor for the Pro tier ($15 vs $20)
Good for scenarios where you want to describe a big task and review the result

Weaknesses

Cascade credits are a real constraint on heavy users
VS Code extension parity is slightly behind Cursor
Requires more upfront thinking about task framing to get good results

Bottom Line

Windsurf is the best IDE-based vibe coding tool for large-scope changes. If your workflow involves frequent big refactors or you are cost-sensitive, Windsurf's Pro tier delivers more per dollar than Cursor for those use cases.

For a direct comparison of Cursor, Windsurf, and Claude Code, see Cursor vs Windsurf vs Claude Code: Best AI Coding Tool for Solopreneurs.

Claude Code

What It Is

Claude Code is Anthropic's CLI-based coding agent. It runs in your terminal, has access to your filesystem, runs shell commands, and executes complex multi-step coding tasks autonomously. No GUI. A command-line tool powered by Claude.

For vibe coding at the highest end, Claude Code is the most capable tool available. You describe a feature or a task, Claude Code reads your entire codebase, plans an approach, writes code, runs tests, fixes errors, and commits the result.

Pricing

Claude Code uses the Anthropic API directly. Pay per token:

Claude Sonnet 4.6: $3 per million input tokens, $15 per million output tokens
Claude Opus 4.6: $15 per million input tokens, $75 per million output tokens

Typical session: a 30-60 minute coding session costs $1-5 on Sonnet, $5-20 on Opus. Heavy users report $50-150/month total.

Best For

Handing off a full feature and expecting working code back
Complex tasks that require multi-step reasoning (authentication flows, database migrations, API integrations)
Developers comfortable in the terminal
Building AI-powered products where Claude Code fits naturally into automated workflows

Strengths

Highest autonomy of any tool on this list
200k token context window loads large codebases in full
Runs your full build, test, and git workflow without switching tools
Pay-per-use pricing scales with actual use, not a flat subscription

Weaknesses

No GUI: not suited for developers who prefer a visual interface
No in-line autocomplete: it is an agent, not a co-pilot
API cost can be high on intensive sessions
Verbose output requires time to review carefully

Bottom Line

Claude Code is the highest-ceiling vibe coding tool. When the task is genuinely complex, it produces better results than any IDE-based alternative. Most serious solopreneurs use Claude Code alongside Cursor or Windsurf: IDE for daily coding flow, Claude Code for heavy-lift tasks they want fully delegated.

Bolt

What It Is

Bolt is a browser-based AI app builder built by StackBlitz. You describe an application in natural language and Bolt generates a full-stack web app running in a live preview inside your browser. No local setup required.

Bolt uses WebContainers, which means the entire development environment runs in the browser tab. You can edit the generated code, install packages, and deploy directly from Bolt.

Pricing

Free: limited daily tokens
Pro ($20/month): 10 million tokens/month, priority support
Teams ($30/user/month): shared workspaces, higher limits

Best For

Non-technical founders who want to build and ship without any local setup
Rapid prototyping where you want a working app in minutes
Demos and MVPs that need to look real quickly
Solopreneurs who want to validate an idea before committing to a full build

Strengths

Zero local setup: works entirely in the browser
Generates full-stack apps (frontend plus backend plus database schema) from a single prompt
Live preview updates as the AI makes changes
Direct deployment to Netlify and other platforms

Weaknesses

Less control over architecture and file structure than IDE-based tools
Generated code quality degrades on complex or highly specific requirements
Token limits on the free and Pro tiers can be hit quickly on large apps
Not designed for working on existing codebases

Bottom Line

Bolt is the best vibe coding tool for non-technical solopreneurs who want to go from idea to deployed prototype with no friction. The tradeoff is control: you get speed and simplicity, but less ability to customize the underlying architecture.

Lovable

What It Is

Lovable (formerly GPT Engineer) is a browser-based AI app builder focused on building full-stack applications. It integrates directly with Supabase for the backend and generates React-based frontends with Tailwind styling.

Like Bolt, Lovable is browser-native. The differentiator is deeper Supabase integration out of the box, which makes it well-suited for solopreneurs building data-driven apps.

Pricing

Free: 5 messages/day
Starter ($20/month): 100 messages/month
Launch ($50/month): 400 messages/month
Scale ($100/month): 1,000 messages/month

Messages map roughly to meaningful edit requests. Complex apps will consume messages faster.

Best For

Solopreneurs building data-driven web apps with Supabase as the backend
Founders who want a polished frontend with minimal design effort
Launching SaaS products quickly where the backend is straightforward

Strengths

Native Supabase integration handles auth, database, and storage automatically
Generated UIs are polished by default (good Tailwind styling out of the box)
GitHub sync lets you export the code and continue in any editor
Strong for CRUD-heavy apps where the data model is the core

Weaknesses

Message-based pricing is less predictable than token-based pricing
Less suited for complex custom logic or non-standard architectures
Free tier is limited compared to competitors

Bottom Line

Lovable is the best vibe coding tool for solopreneurs building a standard SaaS or data app on Supabase. The Supabase integration removes a lot of backend boilerplate and the generated UIs ship well. For apps that fit the mold, it saves significant time.

v0

What It Is

v0 is Vercel's AI UI generator. It converts text prompts and screenshots into React components using Tailwind CSS and shadcn/ui. The output is production-ready component code you copy into your project or deploy directly to Vercel.

v0 is more specialized than the other tools on this list. It is not building full apps. It is building UI components and layouts.

Pricing

Free: 200 credits/month (approximately 10-20 generations)
Pro ($20/month): 5,000 credits/month
Enterprise: custom pricing

Best For

Building polished React UIs fast
Solopreneurs who need a frontend but do not want to do custom CSS work
Developers who already use the Vercel ecosystem (Next.js, shadcn/ui, Tailwind)
Converting design screenshots or wireframes into working component code

Strengths

Best output quality for React UI components of any tool on this list
Generates accessible, styled components using industry-standard libraries
Screenshot-to-code capability is genuinely useful for replicating designs
Direct Vercel deployment requires no configuration

Weaknesses

Narrow scope: UI components only, no backend or business logic
Tied to the React and Tailwind ecosystem
Credits run out quickly if you are generating complex multi-section pages

Bottom Line

v0 is the best vibe coding tool for frontend UI work within the React ecosystem. If your stack is Next.js with Tailwind and shadcn/ui, v0 can generate production-ready component code faster than any other option. Use it alongside Claude Code or Cursor for the full-stack work.

Side-by-Side Comparison

Tool	Interface	Starting Price	Best At	Backend Support	Existing Codebase
Cursor	IDE (VS Code)	$20/month	Daily coding flow	Yes	Yes
Windsurf	IDE (VS Code)	$15/month	Large-scope refactors	Yes	Yes
Claude Code	Terminal / CLI	Pay-per-use	Complex autonomous tasks	Yes	Yes
Bolt	Browser	Free / $20/month	Zero-setup prototypes	Yes	No
Lovable	Browser	Free / $20/month	Supabase-backed apps	Yes (Supabase)	Limited
v0	Browser	Free / $20/month	React UI components	No	Copy/paste

How to Choose

You want to start building immediately with no setup. Use Bolt for the initial prototype, then export to Cursor or Claude Code when you need more control.

You are a developer who wants AI integrated into your daily workflow. Cursor at $20/month is the default choice. Windsurf at $15/month if you do a lot of large-scope changes.

You want to hand off complex tasks completely. Claude Code. Give it a spec, let it run, review the output. This is peak vibe coding for developers comfortable in the terminal.

You are building a data-driven SaaS on Supabase. Lovable handles the boilerplate faster than any other option and the Supabase integration is tight.

You are building a React frontend and want polished components fast. v0 for the UI layer, everything else in Cursor or Claude Code.

You have no coding experience and want to ship something real. Bolt or Lovable. Both are designed for this. Bolt for generic apps, Lovable if your app needs a database from day one.

The Honest Solopreneur Take

Most solopreneurs who are serious about vibe coding end up with two or three tools across different use cases:

An IDE tool (Cursor or Windsurf) as the daily driver
Claude Code for complex, autonomous tasks that need the highest-quality output
v0 or Bolt for fast prototyping and UI generation

This is not redundancy. Each covers a different mode of work. You do not need all three from day one. Start with Cursor or Bolt depending on whether you prefer working in an IDE or a browser. Add the others when you hit their specific use cases.

The wrong move is spending time researching tools instead of using them. Pick one from this list, run it for two weeks, and develop your own opinion. The best vibe coding setup is the one you actually use.

Related Guides

The One-Person Unicorn: What It Takes and What Stack They Run

neo one — Fri, 20 Mar 2026 09:47:37 +0000

This article was originally published on do-nothing.ai. The canonical version is there.

The One-Person Unicorn: What It Takes and What Stack They Run

Sam Altman said one founder with AI could build a billion-dollar company. The tech press covered it as a prediction. It was a description of something already happening.

By 2026, the early examples are in. Here is what they actually look like, what they require, and what the stack behind them runs on.

What the Thesis Is (and Isn't)

The one-person unicorn is not about working 100-hour weeks alone. It's about extreme leverage through AI.

The model:

One human sets strategy, makes judgment calls, maintains relationships
AI handles execution across multiple functions simultaneously
The business scales revenue without scaling headcount at the same rate

This is already happening in narrow domains. Software companies with one founder and AI agents handling support, content, QA, and outbound are reporting $500k-$2M ARR per human.

The unicorn threshold ($1B valuation) requires either:

A high-margin SaaS product with network effects, or
A content or data business with compounding distribution, or
An AI-native service business that can deliver enterprise-quality output at software margins

What It Actually Requires

1. A Defensible Position

AI dramatically lowers execution costs. It does not create moats by itself. The one-person unicorn needs:

A unique insight, dataset, relationship set, or distribution channel
Something that gets harder to replicate as it grows (network effect, proprietary data, brand)
Pricing power that doesn't erode when competitors also get AI

2. Agent-Driven Operations

At one-person unicorn scale, agents are not a convenience. They are how the work happens.

Functions that run on agents:

Customer support triage and first response
Content creation pipeline (research, draft, publish)
Lead qualification and outbound enrichment
Code generation, testing, and deployment
Financial reporting and anomaly detection
Competitive monitoring and market intelligence

Functions that stay human:

Enterprise sales and negotiation
Strategic partnerships
Product direction and prioritization
Public positioning and brand voice

3. Systems That Scale Without You

The one-person unicorn scales because the systems run without the founder's hourly input.

Automated onboarding: user signs up, gets access, gets trained, gets value, no human involved
Automated support: 80%+ of tickets resolved by agents without escalation
Automated content: publishing pipeline runs on schedule, founder reviews weekly
Automated outbound: agents identify and qualify prospects, founder takes the call

4. A High-Margin Business Model

Not every business can be a one-person unicorn. The economics have to support it:

SaaS or API pricing (not hourly services)
Gross margins above 70% (software margins, not agency margins)
Customer acquisition that scales with content, not outbound headcount

The Stack They Run

Based on the pattern of AI-native businesses at this scale in 2026:

Function	Tool
LLM backbone	Claude Opus 4.6 + Haiku 4.5
Agent orchestration	Anthropic Agent SDK or LangGraph
Workflow automation	Make.com or n8n
Product backend	Next.js + Supabase
Payments	Stripe
Support agents	Claude via API with custom system prompt
Content pipeline	Claude + MDX + Vercel
Outbound	Clay + Claude for enrichment and personalization
Analytics	Ahrefs + Posthog
CRM	Airtable or Notion (lightweight)

Total tooling cost at $1M ARR: $2,000-5,000/month (LLM API costs dominate).
At 70%+ gross margins, this is a rounding error.

What the Numbers Look Like

A one-person AI business at $1M ARR with 75% gross margins:

Revenue: $83,000/month
Tooling + AI costs: $3,000-5,000/month
EBITDA (pre-tax): $65,000-75,000/month

At that margin profile, the founder is building both cash flow and enterprise value simultaneously.

The Realistic Timeline

The one-person unicorn path is real but not fast:

Year 1: find product-market fit, validate model, reach $10-50k MRR
Year 2: build automation that removes you from daily execution, reach $100-500k ARR
Year 3+: compounding distribution, repeatable acquisition, $1M+ ARR threshold

The unicorn valuation ($1B) requires either a venture raise at a high multiple or a category-leading revenue trajectory. Most one-person businesses at this scale will exit at $10-50M, which is still life-changing.

What Holds People Back

Underbuilding the agent layer: doing too much by hand, limiting time for the highest-leverage work
Wrong business model: choosing services over software, capping margins
No distribution moat: building a product without a compounding content or network strategy
Scaling prematurely: hiring to solve problems that agents would solve cheaper