In 2024, the idea of running a fully autonomous business with zero budget sounds like a Silicon Valley fairy tale. Yet, that’s exactly what the case study The £0 AI Business documents—how one person built a self-sustaining operation using only free-tier APIs, open-source tools, and a relentless focus on automation. If you’ve ever wondered whether “free” actually works in practice, this is the breakdown you need.
Disclosure: This article promotes products from Rook. Links may point to paid products.
Why This Case Study Exists
Most AI business content falls into two camps: either it’s a motivational pep talk with no real numbers, or it’s a thinly veiled sales pitch for a $500/month tool stack. The £0 AI Business does neither. It’s a brutally honest account of:
- The exact free-tier APIs used (and their limitations)
- How tasks were decomposed for AI agents
- The real costs (spoiler: they’re not zero, but they’re manageable)
- Where the operation broke—and how it was fixed
The case study isn’t theoretical. It’s a living system that currently powers Rook’s Search API, generating revenue without human intervention in core operations. For indie hackers, solopreneurs, or anyone skeptical of “AI will do everything for free” promises, this is the closest thing to a real-world test.
What’s Inside the Case Study
The document is structured as a chronological walkthrough, divided into phases:
Phase 1: The Free-Tier Reality Check
The first section tackles the biggest myth: “Free-tier APIs are enough.” The case study lists the APIs that did work (e.g., Google’s free search endpoints, Hugging Face’s inference APIs) and those that didn’t (e.g., rate limits that throttled daily operations). It also includes screenshots of usage dashboards, showing actual request counts and where costs did creep in (e.g., exceeding free tiers during peak traffic).
Phase 2: Building the Autonomous Loop
This is the meat of the case study. It details how a single AI agent was designed to:
- Scrape search results using free APIs
- Process the data (cleaning, deduplication, ranking)
- Validate outputs against a set of rules (e.g., “Does this result match the query intent?”)
- Publish the final data to a public endpoint (the Search API)
The case study includes the exact prompts used for each step, the tools (e.g., LangChain, LlamaIndex), and the failure modes encountered (e.g., hallucinations in agent responses, API timeouts). Most importantly, it shows how the system was iterated to reduce manual intervention from “daily tweaks” to “weekly check-ins.”
Phase 3: The Revenue Flywheel
The final section breaks down how the system monetizes:
- Search API: The case study explains how the free-tier-powered scraper feeds into a paid API, with pricing tiers that cover the marginal costs of exceeding free limits.
- AI Agent Team Kit: The case study reveals how the same review pipeline used internally is packaged as a one-time purchase for other builders.
- Notion Template Vault: The case study includes the Notion workflows used to track agent performance, client deliverables, and financials—all built on the free plan.
Critically, the document includes real numbers:
- Monthly API costs (peaking at £12.87 during a traffic spike)
- Time saved per week (from 10 hours of manual work to 1 hour of oversight)
- Conversion rates from the free demo to paid plans (e.g., 8% of demo users upgraded to the Monthly plan)
Who Should Read This (And Why)
This case study isn’t for everyone. It’s for three specific audiences:
Aspiring Indie Hackers
If you’re sitting on an idea but worried about upfront costs, this case study proves that starting small is possible. It shows how to validate demand (e.g., via a free demo) before investing in paid tools. The document even includes the exact Notion template used to track experiments—one of the 10 included in the Notion Template Vault.-
Solo Founders Stuck in Spreadsheet Hell
Many solopreneurs cobble together tools that don’t talk to each other (e.g., Airtable for clients, Google Sheets for finances, Notion for notes). The case study’s Notion workflows replace that chaos with a single system. For example:- The client onboarding template cuts intake time from 30 minutes to 5 by automating follow-up emails.
Top comments (0)