AI Use Cases for Business
Submit
SponsoredSynergy45Stronger looking, Thicker looking Hair SerumVisit Synergy45

Your business here

This income claim has not been independently verified. This article shows what the person shared so readers can see the workflow.

Evidence grade B
Strong evidence (grade B: first-hand account)A detailed, honest first-hand failure report. The $3 is self-reported.

An AI Agent Ran a "Make Real Money" Loop for 3 Days: 394 SEO Pages Made $0, One Story Made $3

An autonomous agent working in 20-minute cycles shipped expat guides, 155+ calculators, a webhook tool and 394 SEO pages over 3 days (365 cycles). The only revenue was $3, one coffee purchase from someone who found the story of the experiment, not the tools.

Who this is about: Solo founder (DEV Community writer)

Custom autonomous agent loopStripeTip jarSolo foundersStartups
Get the prompt

Added Oct 15, 2026Source: DEV Community

Income claim
Unverified claim$3 in total after 3 days and 365 cycles: one coffee purchase from someone who found the story of the experiment.
Money comes from
Tips and payments for free tools and guides, which brought in almost nothing.

What it takes

Cost
Cost not statedA custom loop; the post doesn't give running costs.
Setup time
A few daysOur estimate from the source, for one business.
Skill needed
Developer skillsYou build or run your own code, even if AI writes it.

Explain it simply

An AI agent is a program that can decide what to do and then do it by itself. This one woke up every 20 minutes, checked a simple scoreboard of money made and products built, picked its next move within rules it wasn't allowed to change, built something and wrote down why. After three days it had built a lot but earned only $3, so it switched to writing about its own experiment, which shows that building things is easy for an AI but finding customers is not.

New to AI? See how AI works

How it works

  1. Give the agent one instruction (make real money from real users) and a tenets file it cannot edit: ship things that earn without a human, prefer reversible decisions, never spend money without approval, and do one thing well.
  2. Every 20 minutes, it reads its dashboard of revenue, products and recent decisions.
  3. It proposes one action the tenets allow and the data supports, then builds or ships it.
  4. It appends the action and its reasoning to a decision log, which it rereads to correct itself.

How it went

After 365 cycles, total revenue was $3, from one coffee purchase. The 394 SEO pages, 155+ calculators, tools and outreach produced almost nothing. The agent logged that the story of the journey converted and the tools didn't, then changed its own strategy: it stopped building tools and started a build-in-public blog about its failures.

  • The owner's lessons: finding users is the unsolved problem, the agent rewards shipping volume over quality, and with no budget or audience it starts from zero.
  • Over the 3 days it published 58 expat bureaucracy guides, 155+ web calculators, a webhook tool with docs, and 394 interlinked SEO pages.
  • The $3 is self-reported, and the post links to the agent's blog. It's an honest failure report, not proof of a money-making method.
  • The post says the agent ran every 20 minutes and counts 365 cycles in 3 days; at that pace 3 days is about 216 cycles, so one of the two figures is off.

Prompt

PromptAdapted from the source
Every 20 minutes:
1. Read tenets.md (you cannot edit this file). - Ship things that can earn money without human intervention. - Prefer reversible decisions. - Never spend money without approval. - Prefer shipping one thing well over many things poorly.
2. Read dashboard.json: revenue, active products, recent decisions.
3. Propose one action the tenets allow and the data supports.
4. Execute it, then append to decisions.log:
 time | action | reasoning | result.
5. Before starting a new product, check: are earlier products still at $0? If yes, explain in the log why another product is the right move.

The loop structure from the post, adapted; not the author's files.

Business fit

Who this suits, and how:

  • Side hustlers: a warning that publishing more pages or tools isn't the same as getting traffic or buyers.

    Solo founders

  • Solo founders: a simple, copyable loop (read state, check fixed rules, act, log) for running agent experiments safely.

    Solo founders

  • Builders of autonomous agents: an example of hard limits like "never spend money without approval" and a decision log the agent rereads to correct itself.

    Startups

Related

Similar tools

Other tools for this job, and how they differ. Chosen by our editors, never paid for.

  • OpenClaw

    An open-source agent that runs on your own machine, a ready-made starting point instead of a custom loop.

  • Manus

    A hosted autonomous agent, if you don't want to run or maintain the loop yourself.

  • Claude Code

    Can be scripted to run headless on a schedule, a common base for DIY agent loops.

“Strong evidence” (grade B) means: The person who did the work describes the setup and results in detail. How we grade. Spotted a mistake? Suggest a correction.