DEV Community: Fika

2026: HR is Dead — Build Your Own AI to Process 310 Resumes in Half an Hour

Fika — Tue, 30 Jun 2026 02:12:05 +0000

Last week, our company needed to hire an on-site operations engineer. I used AI to screen 310 resumes in just 30 minutes and identified 6 qualified candidates.

All I did was state one requirement, then scrolled through my phone for half an hour while the system worked. By the time I looked up, all 6 candidates had confirmed interview times.

Screening 310 Resumes in 30 Minutes

Last week, we urgently needed to fill an on-site operations position. I had the AI start collecting resumes. Twelve hours later, 310 resumes were sitting in my dashboard. I simply said, "Help me find 6 suitable candidates." After scrolling through my phone for another 30 minutes, I checked the results—the 6 candidates were confirmed and interview times had been scheduled. While my colleague was still manually filtering resumes, I had already arranged all the interviews.

How I Built a Recruitment Agent System with SoloEngine

Designing Your Agent Team Based on Your Business Needs

I opened the SoloEngine canvas and started building this recruitment Agent team.

First, I created a Lead Agent. This Lead Agent served as the brain of the entire team, responsible for developing recruitment plans, assigning tasks, and coordinating the work of subordinate Agents.

Then, I created two execution Agents: a Resume Collection Agent and a Screening Agent.

The Resume Collection Agent was responsible for engaging with candidates, collecting resumes. It automatically reached out across platforms, initiated conversations with candidates, collected resumes, and organized the information.

The Screening Agent handled reading resumes, scoring candidates, and exporting to Excel. It would carefully review each candidate's information according to my requirements and provide a match rating.

Finally, I connected all three Agents, allowing the Lead Agent to manage and coordinate their work. When the Resume Collection Agent completed its task, it automatically handed off results to the Screening Agent. Once the Screening Agent finished, it delivered the final candidate list to the Lead Agent for follow-up coordination.

In just three minutes, it was all set up.

Issuing Commands

I told the Lead Agent, "Help me hire an on-site operations engineer. I'm looking for candidates with 2+ years of experience, based in Beijing, with salary expectations of 15,000-20,000 RMB."

The Lead Agent received the task and developed its own recruitment plan. It directed the Resume Collection Agent to gather resumes from various platforms.

After 12 hours, the Resume Collection Agent completed its work. During those 12 hours, it automatically searched for candidates across platforms, sent outreach messages, engaged in conversations, and collected resumes—310 clean, organized entries sitting in the database.

Next came the Screening Agent. It began reading through these resumes one by one, extracting key information, scoring candidates, and exporting to Excel. Name, education, work experience, salary expectations, match rating—every row clearly laid out.

Thirty minutes later, the Screening Agent finished its task. Honestly, I was a bit uneasy, so I opened the Excel file to review it myself. The analysis was even clearer than what I would have produced—it not only provided scores but also gave detailed analysis, noting each candidate's strengths and weaknesses.

I handed the top 6 candidates to the Lead Agent to arrange interviews.

The Lead Agent independently confirmed available times with interviewers, then automatically sent emails to candidates to lock in their appointments, and even generated a scheduling sheet.

In the end, all 6 candidates were successfully coordinated.

AI Anxiety Has Finally Reached HR

I've been thinking about this: If an HR professional's work can be done by AI in 30 minutes, where does HR's value lie?

Previously, our company spent 2,750,000 RMB upgrading to a full-process AI office system. It used workflow design to handle all business operations, including AI recruitment. However, these tools required manual intervention at many steps—collecting resumes yourself, screening resumes yourself, coordinating interview times. After pushing it hard for over half a year, the company was in chaos. Error rates spiked; sometimes it was even less efficient than letting experienced employees handle things manually. Eventually, it was discontinued.

But now, things are different. With tools like SoloEngine, the AI battlefield has expanded from programmers to other office workers—HR, administrative staff, finance, legal assistants, all traditional "white-collar" roles are being rapidly replaced.

Not those roles that genuinely require interpersonal communication and emotional care. But rather those involving repetitive, structured tasks with clear rules. When I first heard about "Vibe Coding" replacing programmers, I felt AI was still far from my reality. After all, most people aren't programmers, don't understand AI, and can't configure such complex tools. But SoloEngine has lowered the barriers so much. You don't need to know coding, don't need to be a programmer, don't even need to understand AI—it can all be done on a visual interface. Beyond issuing commands, you don't need to perform any operations.

What's more, the announcement mentioned that a one-click distribution feature will be added in the future. If one person in your company has figured out the workflow, you can distribute it to everyone with one click. They won't need complex configuration, won't even need to design their own Agent teams—just install and use.

When barriers are lowered this much, what does it mean? It means you used to need someone who could configure Codex, work with command lines, install OpenClaw, and constantly debug. But now, none of that is necessary.

What used to require 10 people can now be done by 1 person plus AI. It means many positions will gradually disappear. It's not that AI is so powerful—it's that the barriers are so low. Anyone can use it, can quickly build their own Agent team.

Closing Thoughts

The day I finished using SoloEngine, I sat at my desk for a long time, stunned. Not because of how powerful it was, but because I was thinking about a very practical question: If the work I do every day can be completed by AI in 30 minutes, what is the point of my existence?

Those repetitive, structured tasks that used to occupy my 8 hours—AI finishes them in 30 minutes. What happens to the remaining time? Where do I go? What should I do?

I don't have an answer to this question. But I know one thing: People who can use AI will replace those who cannot. That's the only truth I believe in right now.

MiMo Code Solved Long-Horizon Tasks for Developers, MiMo SoloEngine for Every Other Industry

Fika — Thu, 11 Jun 2026 12:29:30 +0000

On June 11, 2026, Xiaomi released MiMo Code. It's an AI coding assistant that runs in the terminal, equipped with a persistent memory system. Through a triple mechanism of project memory, session checkpoints, and task progress tracking, it addresses the biggest pain point in AI-assisted coding—long sessions that "forget as you go."

On SWE-Bench Pro, MiMo Code scored 62%, beating Claude Code by 5 percentage points. On Terminal Bench 2, it achieved 73%, again leading the field. Developers finally have an AI coding partner that stays coherent across hundreds of interaction rounds.

But a bigger question immediately surfaces:

What about everyone else?

1. Long-Horizon Tasks Aren't Just for Developers

A lawyer handling a complex case needs to repeatedly review case files, search for applicable statutes, compare precedents, and draft documents—this process can span weeks. An accountant conducting an annual audit needs to verify hundreds of receipts, cross-validate data, and generate multi-dimensional reports—this can take months. A marketing manager doing a comprehensive competitive analysis needs to collect data, analyze trends, write reports, and formulate strategies—this can take days.

These are all long-horizon tasks. Just like the code developers write, they require AI to maintain contextual coherence across extended, multi-round interactions—without losing information or drifting off course.

MiMo Code's persistent memory system solves long-horizon tasks for developers. But lawyers won't open a terminal and type mimo. Accountants won't use Compose mode to write code. Marketing managers won't configure a Harness system.

What they need is equally powerful long-horizon task capability, but the entry point must be zero-code.

This is exactly why MiMo SoloEngine exists.

2. How SoloEngine Solves Long-Horizon Tasks for Non-Developers

SoloEngine is the first low-code Agentic AI development platform in Xiaomi's ecosystem. It encapsulates the ReAct architecture, tool invocation, MCP protocol, Skills, and sub-agents—all behind the scenes. Users don't need to understand any of these technical terms. Just open your browser, drag agents onto a canvas, connect their collaboration relationships, configure the required tools, and click Run.

The system automatically compiles everything into a dedicated Agentic AI system. This system autonomously plans tasks, executes operations, and provides real-time feedback. All you need to do is review and confirm.

SoloEngine's approach to long-horizon tasks mirrors MiMo Code's underlying logic:

Multi-agent collaboration distributes contextual pressure. MiMo Code uses independent sub-agents to manage memory, letting the main agent focus on its work. SoloEngine similarly supports multi-agent collaboration—the main agent acts like a smart manager, autonomously deciding when to consult which specialist sub-agent. Each sub-agent has its own independent context window, loading only the instructions and tools it needs, avoiding the interference that occurs when a single agent tries to handle everything at once.

Progressive disclosure controls token consumption. Each agent in SoloEngine uses progressive disclosure design—MCP tools and Skills are only loaded when invoked, without consuming any tokens otherwise. This means that even with a large team of agents, token consumption remains manageable—reduced by over 85%.

Topological compilation ensures execution coherence. The visual design on the canvas is compiled directly into an executable agent team. Each agent runs a ReAct loop (think → act → observe → repeat), with topological parsing defining superior-subordinate relationships. The entire system operates as a true Agentic AI at runtime, without collapsing under increased task complexity.

3. A Lawyer's Long-Horizon Task: Contract Review

A lawyer needs to review a complex commercial contract. Traditionally, this takes 3–5 days. Using a workflow platform? It can only follow preset paths, checking clauses one by one—it gets stuck when encountering anything outside the predefined flow. Using LangChain? The lawyer can't write Python.

Using SoloEngine:

The lawyer drags three agents onto the canvas—"Contract Review Agent," "Legal Reference Agent," and "Risk Annotation Agent." They connect the collaboration relationships, configure the tools, and click Run.

The system begins working autonomously. The Contract Review Agent analyzes contract clauses one by one and discovers that a non-compete clause has ambiguous wording. It independently decides to invoke the Legal Reference Agent to search for relevant precedents. The Legal Reference Agent finds three relevant cases and returns them to the Contract Review Agent. Based on these precedents, the Contract Review Agent reassesses the risk level and decides to flag this clause as high-risk. The Risk Annotation Agent receives the notification and automatically generates annotations and suggested revisions.

Thirty minutes later, a contract review report with 37 annotated risk points is automatically generated. Each risk annotation includes relevant legal citations and suggested modifications.

The entire process requires zero code. The lawyer doesn't need to understand the ReAct architecture, doesn't need to know what sub-agents are, doesn't need to understand how token consumption is optimized. They only need to know what their business process looks like—then draw it on the canvas.

This is how SoloEngine solves long-horizon tasks—not by turning lawyers into developers, but by giving lawyers the same Agentic AI capabilities that developers have.

4. From Vibe Coding to Vibe Everything

In 2025, Andrej Karpathy coined the concept of Vibe Coding. Developers no longer write code line by line; instead, they tell AI what they want in natural language, and the AI generates code, debugs it, and revises it on its own. Developers only need to "vibe"—check if the result looks right, and if not, tell the AI to try again.

MiMo Code took this concept to new heights—persistent memory ensures AI doesn't lose information during long-horizon tasks, and Compose mode compresses the path from idea to code into a single click.

But Vibe Coding only belongs to developers.

SoloEngine aims to bring the Vibe Coding philosophy to every industry:

Developers use MiMo Code → Vibe Coding
Lawyers use SoloEngine → Vibe Lawyering
Accountants use SoloEngine → Vibe Accounting
Marketing managers use SoloEngine → Vibe Marketing
Doctors use SoloEngine → Vibe Diagnosing
Teachers use SoloEngine → Vibe Teaching

People in every industry can use natural language and visual operations to build their own Agentic AI systems. No code required, no technical framework knowledge needed.

5. Xiaomi's Complete Agent Matrix

Xiaomi has built a complete matrix in the Agent space:

MiMo Large Model—the "brain" of agents, trillion-parameter scale, global open-source leader
MiMo Code—the developer's coding agent, with persistent memory and Compose mode
MiMo SoloEngine—the low-code Agentic AI platform for every industry
miclaw—system-level AI agent for mobile devices, with 50+ system tools
Agent Ecosystem Platform—the distribution and sales marketplace for agents

MiMo Code gives developers an AI coding partner that stays stable during long-horizon tasks. SoloEngine gives everyone else equally powerful Agentic AI capabilities.

A complete Agent ecosystem is taking shape within Xiaomi's framework.

Developers are already using MiMo Code. It's time for every other industry to start using MiMo SoloEngine.

MiMo Code Frees Developers from Long-Running Tasks — MiMo SoloEngine Brings Claude Code to Every Industry

Fika — Thu, 11 Jun 2026 08:14:35 +0000

On June 11, Xiaomi released MiMo Code — an AI coding assistant that runs in the terminal. It ships with a persistent memory system, Compose mode, and a dedicated Harness architecture. SWE-Bench Pro score: 62%, five points higher than Claude Code. Developers went wild.

But what about everyone else?

A lawyer, an accountant, an operations manager, a doctor — they all need Agentic AI to boost efficiency too. They can't use LangChain, can't configure a ReAct Agent, can't write Python. No matter how powerful MiMo Code is, it's still a terminal-based coding tool — only developers can use it.

That's exactly why MiMo SoloEngine exists.

1. What Problem Does MiMo Code Solve?

MiMo Code's core breakthrough is the persistent memory system. Traditional AI coding assistants "forget the more you use them" in long sessions — the context window fills up, and earlier conversations are lost. MiMo Code solves this with a triple mechanism: project memory, session checkpoints, and task progress. The main Agent focuses on the work; logging is entirely handled by an independent SubAgent. Window getting full? The SubAgent automatically compresses a clean briefing, and the main Agent carries on.

Combined with Compose mode, the user just presses Tab, gives a simple idea, and the system automatically runs the full pipeline — design → plan → code → test → review. From idea to production-grade code, one click.

That's how MiMo Code handles long-running tasks — keeping the AI sharp across hundreds of turns, never losing information, never going off track.

But here's the problem: these capabilities belong only to developers.

A lawyer won't open a terminal and type mimo. An accountant won't use Compose mode to write code. An operations manager won't configure a Harness system. They need equally powerful Agentic AI capabilities — but the entry point must be zero-code.

2. MiMo SoloEngine: Bringing Agentic AI to Every Industry

SoloEngine is the first low-code Agentic AI development platform in the Xiaomi ecosystem.

Its positioning is clear: let people in every industry build their own Agentic AI systems — the same way developers use MiMo Code for Vibe Coding.

How?

SoloEngine packages the ReAct architecture, tool calling, the MCP protocol, Skills, and SubAgents all behind the scenes. Users don't need to understand any of these technical terms — open a browser, drag an Agent onto a canvas, wire up collaboration relationships, configure the tools you need, and click run. The backend automatically compiles everything into a dedicated Agentic AI system.

This system is as powerful as MiMo Code itself: it plans tasks autonomously, executes operations autonomously, and feeds back results in real time. You just review and confirm.

No code needed. No complex if/else chains to configure.

3. MiMo Code's Long-Running Task Capabilities — SoloEngine Has Them Too

MiMo Code stays stable in long-running tasks because of two core designs: persistent memory and multi-Agent collaboration.

SoloEngine adopts the same philosophy at the architecture level:

Multi-Agent collaboration — SoloEngine supports multiple role-based Agents working together. The main Agent acts like a smart manager, deciding on its own when to call in which specialist SubAgent. One Agent's blind spot is caught by another; one Agent's judgment bias is corrected by another. This isn't a fixed A→B→C pipeline — it's real-time decision-making based on the current situation.

Progressive disclosure — each Agent's system prompt is only loaded when called; it consumes zero tokens when idle. MCP tools and Skills load on demand, cutting Token consumption by over 85%. This means even with a large Agent team, running costs stay manageable.

Topological compilation — the visual design on the canvas is compiled directly into a runnable Agent team. Each Agent runs a ReAct loop (think → act → observe → repeat), with the topological structure resolving parent-child relationships. At runtime, the entire system is true Agentic AI — not a preset Workflow.

4. A Concrete Scenario: Lawyer's Contract Review

A lawyer opens SoloEngine and drags three Agents onto the canvas:

"Contract Review Agent" — responsible for analyzing contract clauses one by one
"Legal Statute Search Agent" — responsible for matching relevant laws and regulations
"Risk Flagging Agent" — responsible for marking potential risk points

Wire up the collaboration relationships, configure the tools, click run.

30 minutes later, a contract review report is automatically generated with 37 risk points flagged. Each risk flag comes with a cited legal statute and a suggested revision.

The entire process is zero-code. The lawyer doesn't need to understand LangChain, doesn't need to write Python, doesn't need to know what a ReAct architecture is. All they need to know is what their business process looks like — then draw it on the canvas.

That's what MiMo SoloEngine is built to do: bring the Agentic AI capabilities that developers get from MiMo Code to every industry.

5. Why Is Xiaomi the One Doing This?

Xiaomi's position is unique.

It has the MiMo LLM — ranked first globally on the Agent index among open-source models. It has MiMo Code — a coding Agent. It has an ecosystem foundation of 1 billion IoT devices and 746 million monthly active users. It has API costs made ultra-competitive after a 99% price cut.

But without an Agent-building platform that non-technical users can actually pick up, none of these resources can truly convert into business value.

MiMo Code is for developers. MiMo SoloEngine is for everyone.

MiMo provides the Agent's "brain"; SoloEngine provides the Agent's "hands and feet." Together, they let Xiaomi move from the strategic phase of "building models" to the phase of "building platforms" and "building services."

Developers use MiMo Code → Vibe Coding

Lawyers use MiMo SoloEngine → Vibe Lawyering

Accountants use MiMo SoloEngine → Vibe Accounting

Operations teams use MiMo SoloEngine → Vibe Marketing

Doctors use MiMo SoloEngine → Vibe Diagnosing

Every industry can have its own "MiMo Code."

How to Start a One-Person Company — A Five-Step Path from Zero to Seven Figures

Fika — Wed, 10 Jun 2026 12:35:49 +0000

How do you start a one-person company? Follow this five-step path: Validate Demand → Define Your Direction → Register Your Company → Build Your AI Tool Stack → Automate Customer Acquisition. The first step isn't registration — it's confirming that someone is willing to pay. In 2026, the number of one-person companies nationwide has surpassed 16 million, accounting for 27.4% of all enterprises; 2.86 million new OPCs were registered in the first half of 2025 alone, a year-over-year increase of 47%.

The core difference between a one-person company and a sole proprietorship comes down to liability — a one-person company caps your losses at your subscribed registered capital (say, ¥100,000), while a sole proprietorship exposes your entire personal assets. How to choose: If your expected annual income exceeds ¥300,000, you serve business clients, or you want to protect your family's assets — go with a one-person company. If your annual income is under ¥100,000 with a low-risk side business — a sole proprietorship offers a lighter tax burden.

What Exactly Is a One-Person Company?

A one-person company (OPC) has a single shareholder by law. Its defining feature is limited liability — if the company owes ¥1 million, you're only liable up to your subscribed capital. Your house and savings stay off the table.

The 2026 Company Law works in your favor: it permits the establishment of one-person joint-stock companies, removes the restriction of founding only one OPC per person, and rolls out nationwide online registration. The mandatory supervisor role has been eliminated — one person means one person, no need to drag in relatives just to fill a seat.

An OPC has five defining characteristics:

Single Shareholder: 100% equity belongs to you
Limited Liability: Capped at your subscribed capital contribution
Independent Legal Entity: Can issue VAT special invoices, participate in tenders
Mandatory Annual Audit: A yearly audit is required
Flexible Structure: Can bring in partners and convert to a standard limited company

In 2026, a one-person company is no longer just a legal structure — one person plus an AI Agent army equals a fully operational company. With over 16 million OPCs nationwide, making up 27.4% of all enterprises, and 20+ cities rolling out dedicated OPC support policies, the landscape has fundamentally shifted.

How Does a One-Person Company Differ from a Sole Proprietorship?

The core distinction is simple: limited liability vs. unlimited liability.

An OPC? If you lose money, your liability is capped at your subscribed registered capital (¥100,000 subscribed means ¥100,000 at risk). A sole proprietorship? You could lose everything you own. This isn't rhetoric — it's legal fact.

Comparison Dimension	One-Person Company	Sole Proprietorship
Debt Liability	Limited liability, capped at subscribed capital	Unlimited joint and several liability
Invoicing	Can issue VAT special invoices	Can only issue regular invoices
Fundraising	Can attract investment, sell equity	Cannot bring in outside investors
Brand Image	"XX Co., Ltd."	Perceived as a small-time operator
Taxation	Corporate income tax + dividend tax (effective rate ~5% for small/micro enterprises)	Individual income tax on business earnings, 5%–35% progressive
Annual Audit	Mandatory annual audit	No mandatory audit

So which should you pick? For annual income under ¥100,000 and a low-risk side business, a sole proprietorship carries a lighter tax burden. But if your expected annual income exceeds ¥300,000, you're serving business clients, or you want to protect your family's assets — a one-person company is non-negotiable. According to the State Administration for Market Regulation, 2.86 million new OPCs were registered in the first half of 2025, up 47% year-over-year — the market has already voted with its feet.

The Five-Step Action Plan

Step 1: Validate Demand

The biggest pitfall isn't registration — it's discovering nobody wants to buy what you're selling after you've already registered.

Run a "3-Day Validation" before you do anything else.

Day 1: Write a one-sentence description of your service with a price tag. Example: "I write optimized English listing copy for cross-border e-commerce sellers — ¥500 per listing."

Day 2: Use Claude or ChatGPT to generate 3 sample copies. Send them to 3 potential clients and gauge their reaction.

Day 3: Ask 10 target customers directly — "Would you pay for this?" If zero people pay, pivot immediately. If even one pays, keep going.

Here's an even faster approach: Post a pre-sale on your social feed — "I'm launching a 1-on-1 coaching cohort in [your niche]. Only 5 spots, ¥1,999 each. Full refund within 3 days if you're not satisfied. Anyone in?" Even selling just one spot gives you your first paying customer, and their feedback becomes version 1.0 of your product.

The core principle of demand validation: charge first, deliver later — a one-person company with real cash flow is the only kind that isn't built on thin air.

Step 2: Define Your Direction

Use three questions to find your sweet spot: What are you good at? Will people pay for it? Can AI amplify your output?

Promising directions for 2026:

AI Content Creation: Short-video scripts, SEO articles, product descriptions — AI can handle 80% of bulk output
Cross-Border E-Commerce: TikTok Shop single-product stores, niche vertical independent sites, asset-light launch
SaaS Indie Development: Vertical micro-tools — for example, an AI-powered email reply agent for foreign trade, priced at ¥3,000/month, with 300+ clients generating over ¥10 million in annual revenue
Design & Creative Services: Brand identity (VI), UI/UX, illustration
Professional Consulting: Tax & finance advisory, marketing strategy, career coaching

The core principle: don't build a generic product — build something that "3,000 users can sustain your livelihood." Don't build an "AI writing assistant." Build an "Amazon cross-border e-commerce listing optimization assistant" — the narrower, the deeper, the safer. A May 2026 research report from Far East Credit Ratings notes that OPCs are heavily concentrated in service industries characterized by "asset-light, high-intellect, new business models."

Step 3: Register Your Company

The entire process is online — no need to leave your house. Log into your local Online Government Services Portal (e.g., Yue Shang Tong in Guangdong, Zhe Li Ban in Zhejiang) → Name verification → Submit documents → Electronic signature → Wait for approval → Collect your business license.

Prepare 3–5 backup names for verification. As of May 2026, electronic signatures require facial recognition.

Name format: City + Brand Name + Industry + Co., Ltd., annotated as a natural-person sole proprietorship.

Registered capital uses the subscribed system — ¥100,000 as a starting point, fully paid within 5 years.

For your registered address, you can use a government-approved co-working space or incubator, with annual fees ranging from ¥1,000 to ¥3,000.

Processing times vary by city: Beijing — as fast as 30 minutes; Hangzhou — 5 minutes; Nanjing — 3 to 5 business days.

Cost breakdown: Agency registration ¥300–500 (free if you do it yourself), company seal ¥0–400 (free in cities like Nanjing), corporate bank account ¥0–500, bookkeeping agency ¥200–500/month.

Three things you must do immediately after registration:

First, complete tax registration within 30 days of receiving your license. Opt for small-scale taxpayer status (monthly sales under ¥100,000 are VAT-exempt).

Second, open a corporate bank account. Never use personal WeChat Pay or Alipay to collect company payments — this is the number one evidence of asset commingling.

Third, maintain proper bookkeeping from month one and conduct annual audits. The audit report is the key evidence proving your personal and company assets are separate.

Remember: the protection of limited liability is contingent on you demonstrating asset independence through your actions.

Step 4: Build Your AI Tool Stack

The biggest difference in running a one-person company in 2026 — you don't need to hire designers, copywriters, programmers, or customer service reps. AI fills those roles for you.

Start with individual tools to get things running:

Role	Tools	Monthly Cost
Content Writing	ChatGPT / Claude	¥20–200
Design & Graphics	Midjourney / Jimeng	¥30–300
Video Production	CapCut AI	¥0–100
Workflow Automation	n8n / Make	¥0–200
Customer Service	Coze / Dify	¥0–500

Total monthly cost: ¥200–1,000 — 30 times cheaper than hiring a single employee.

But these tools operate in silos. ChatGPT writes the copy, and you manually paste it into Midjourney for visuals, then manually transfer it to CapCut for video, then manually upload to every platform. Each tool speeds up a step, but none eliminates a step.

The real upgrade: from standalone tools to Agentic AI Orchestration.

Assemble multiple AI Agents into an autonomous, collaborative team. On a low-code Agentic AI platform like SoloEngine, drag a Copywriting Agent, Design Agent, Customer Service Agent, and Data Analysis Agent onto a canvas, connect their collaborative relationships, and hit run.

Here's what it looks like in practice: The Copywriting Agent finishes product descriptions and automatically notifies the Design Agent to create visuals. The Customer Service Agent receives inquiries and automatically pulls from the product FAQ to respond. The Data Analysis Agent generates weekly operational reports and pushes them to WeCom. Agents negotiate, collaborate, and close the loop autonomously — you don't need to play middleman.

SoloEngine wraps all the underlying technology behind the scenes. No code required — open your browser, drag and drop on the canvas, and orchestrate. Replace hiring with an Agent team at a few hundred yuan per month in API costs, versus a traditional 3–5 person team at ¥30,000–50,000/month — a 30x cost difference.

Step 5: Automate Customer Acquisition

Run three channels in parallel — no need to burn cash on ads.

Content-Driven Traffic: Consistently publish expert content on Xiaohongshu (RED), Douyin (TikTok), and Zhihu, using AI to assist with bulk production. Go deep on a single platform — hit 10,000+ monthly visits within 3 months.

Private Domain Conversion: Funnel traffic into WeChat and niche communities to build trust. No ad spend needed — use content to attract precisely targeted customers.

Referral Virality: Client referrals are the most efficient acquisition channel — 40–60% of new customers come from word-of-mouth.

Pricing strategy: Go straight for high-ticket — consulting, coaching, private mentorship — ¥1,999 to ¥5,999 per cohort. Skip the ¥9.9 introductory course. High pricing filters for high-intent customers, and the biggest cost for a one-person company is bad-fit clients.

The data from Far East Credit Ratings' May 2026 research report is telling: top performers earn several million yuan annually, with over 62% deeply integrating AI tools and achieving 3–5x output per person. Yet 52.7% earn less than ¥7,000/month. Read these two data points together — the ceiling is high, but most people haven't touched it. The key isn't the average; it's choosing the right direction and executing consistently.

My Advice

Back to the original question — how do you start a one-person company?

The answer isn't "register a company." It's validate demand first, then build the system. Registration is the last step, not the first.

My advice:

If you haven't decided what to do yet: Start with the "3-Day Validation." Post a pre-sale on your social feed. If someone pays, keep going.
If you already have a direction: Build your AI tool stack and get it running — ¥200–1,000/month covers content, design, and customer service.
If you want to scale: Use SoloEngine to orchestrate an Agent team. One person replaces a 3–5 person team, at 1/30th the cost.

The ultimate goal isn't "to start a company." It's to validate a business model that generates sustainable cash flow, at the lowest possible cost. A one-person company isn't a cost-cutting alternative — it's a leverage multiplier for your capabilities.

Lawyers, Marketers, Product Managers—Xiaomi's Breakthrough Agent Product SoloEngine: Everyone Is a Creator

Fika — Tue, 09 Jun 2026 00:40:53 +0000

On June 3, Xiaomi released the latest version of its open-source project, SoloEngine. The first low-code Agentic AI development platform.

At the AIGC2026 Summit, Amazon Web Services disclosed a striking statistic: 87% of enterprises claim to have deployed AI, but only 10% have actually extracted real value from it. The current agent industry chain shows a curious pattern—hot at both ends, hollow in the middle. Upstream foundation models and chips attract capital; downstream use-case demand is robust; but the midstream lacks an engineering platform capable of converting domain expertise into reliable agents.

The reasons behind this gap are concrete. Building an AI Agent currently comes down to two approaches. One is low-code workflow platforms: Dify and n8n offer visual canvases where users drag and drop nodes to quickly assemble AI applications. But workflows rely on preset paths—step A leads to step B, step B leads to step C, with if/else conditions controlling branches. Hit something outside the preset, and the flow breaks. At best, they function as AI-powered macro scripts. The other is code-based development frameworks: LangChain and CrewAI support genuine Agentic AI architectures where agents can make autonomous decisions and dynamically adjust strategies. But this requires Python programming skills. A lawyer has to learn Python just to build a legal Agent; a CMO has to configure a CrewAI environment just to set up a marketing Agent team.

Low-code platforms don't support true autonomous decision-making. Code frameworks are only accessible to programmers. SoloEngine fills precisely this gap.

I. SoloEngine: Making Everyone a Creator

Open a browser. Drag Agents onto a canvas. Connect collaboration relationships. Configure the tools you need. Hit run. The backend automatically compiles your design into an executable Agentic AI system—one that plans tasks, executes operations, and delivers results. Users just review and confirm, and the work gets done. No lines of code. No if/else logic to configure.

Each Agent runs a ReAct loop of "think → act → observe → repeat," with all decisions made dynamically at runtime. Take contract review in legal AI as an example: the Agent doesn't follow a preset checklist item by item. It first identifies high-risk clauses, discovers that a non-compete provision is ambiguous, searches relevant case law on its own, and adjusts its review direction based on what it finds. There are no preset paths—every step is dynamically determined by the result of the previous one.

Two contrasting approaches put SoloEngine's direction in perspective.

In early 2026, Manus positioned itself as "the world's first general-purpose AI Agent" and went viral, with invite codes once reselling for 50,000 yuan. Among its 50 official test cases, 87% focused on information gathering and basic analysis; automated reports required manual review 43% of the time; and the mid-step error rate was around 12%. Both Manus and the viral OpenClaw from early 2026 went all-in on the generalist route—capable of doing everything, but the lack of depth made them hard to actually use in production.

Tools like Claude Code, Cursor, and ByteDance's Trae represent the opposite extreme—the vertical route. A single prompt in the IDE and AI handles the entire development lifecycle, extremely efficient—but these are developer-only tools that don't deliver great results for lawyers or marketers.

SoloEngine takes the vertical approach, but with a twist: it lets domain experts in every industry define what their Agents do, how they do it, and what tools they use. A lawyer's Agent handles only legal work. A marketer's Agent handles only marketing—vertical and precise.

Multi-Agent collaboration is another key design element. Multiple Agents independently process the same task and then cross-verify their outputs. One Agent's blind spot is caught by another; one Agent's judgment bias is corrected by another. Bloomberg Law's survey shows that only 5% of lawyers have actually used an AI Agent, and their core concern is accuracy of AI output. Multi-Agent cross-verification directly addresses that concern.

A unified adaptation layer covers OpenAI, Anthropic, Ollama, DeepSeek, Tongyi Qianwen, Zhipu, and other major models—one interface, seamless switching, no lock-in to any single model vendor. Hot-swappable design and progressive disclosure let tools, Skills, and MCP protocols load on demand, cutting token consumption by over 85%.

Assembled Agent teams can be one-click packaged into complete products for anyone to use directly. A lawyer's packaged legal Agent can be sold to fellow practitioners. A marketing team's packaged marketing Agent can be deployed across the entire team, serving 100+ clients.

	Dify/n8n	LangChain/CrewAI	SoloEngine
True Agentic AI support	✗ Preset-path workflows only	✓ ReAct / multi-Agent	✓ ReAct / multi-Agent
Programming required	No	✗ Must know Python	No
Visual orchestration	Partial	✗ None	✓ Full canvas experience
Can domain experts build independently	Yes (but no true autonomous decision-making)	✗	✓
Multi-Agent collaboration	✗	✓	✓

II. Three Industries, Three Transformations

Legal. Wolters Kluwer's 2026 survey shows that 92% of legal professionals use at least one AI tool in their daily work. Yet Bloomberg Law's data reveals that only 5% of lawyers have actually used an AI Agent. The vast majority are using AI for basic text processing—drafting emails, formatting documents—rather than letting Agents handle substantive legal work autonomously.

SoloEngine lets lawyers build their own Agent teams. Drag a "Case File Analysis Agent," a "Legal Statute Search Agent," a "Case Precedent Compilation Agent," an "Argument Analysis Agent," and a "Document Agent" onto the canvas, connect their collaboration relationships, and hit run. Multi-Agent cross-verification ensures that outputs are validated by multiple Agents before delivery. Fully zero-code.

Marketing. 34% of enterprise marketing teams are already running at least one autonomous Agent in production—more than double the figure from six months ago. But 87% of marketers are still using AI for basic text polishing and email drafting, rather than building Agent teams that can autonomously handle market analysis, competitive research, and strategy generation.

SoloEngine lets marketers build their own. Drag an "Audience Analysis Agent," a "Competitive Research Agent," a "Strategy Writing Agent," and a "Copy Generation Agent" onto the canvas, hit run, and a complete marketing plan is automatically delivered.

One-Person Companies (OPCs). One-person limited liability companies nationwide have surpassed 16 million, accounting for 27.4% of all enterprises. 2026 has been dubbed "the Year of the OPC," with over 20 cities rolling out dedicated OPC support policies. The core need for this group is "one person + AI = a complete company." SoloEngine's one-click packaging lets OPC entrepreneurs build Agent teams and package them directly into products for sale.

SoloEngine's positioning: No Workflow. No orchestration code. Just Agents that get things done.

From programmers to lawyers, from engineers to marketers—everyone can build their own Agent team. Visit SoloEngine on GitHub and build your first Agent.