In February, I posted what I’d built in a single month with AI tools. People thought it was impressive.
That was one month. Here’s the full quarter.
January through March 2026. One person. AI subscriptions under $300/month. No employees. No contractors. No outsourcing.
What Shipped
Websites and Content
- 3 complete webbooks published — 32 chapters of neuroscience-based therapy content, 36 chapters of theology, and 12 chapters of creative cosmology. Each one a full standalone reading experience with navigation, table of contents, and responsive design.
- A full Library landing page — 56 curated posts organized across 6 thematic sections, designed and deployed in a single session.
- 19 sermons batch-published to WordPress — uploaded, categorized, and live in one afternoon using parallel AI agents and the WordPress REST API.
- An interactive Pastoral Simulator — a web application that simulates real pastoral counseling scenarios. Built, tested, and deployed to production.
- 15+ browser-based games built and deployed — playable applications from concept to live URL.
AI and Voice
- Voice-cloned IVR phone system — 26 custom audio files generated from a 57-second voice sample using open-source models running on my desktop CPU. Connected to Twilio. Live on our business phone line. Monthly cost: ~$40. Equivalent outsourced answering service: $200-500/month.
- AI companion with lip sync, expression blending, and head-nod tracking — a 3D animated character that talks, emotes, and responds to conversation in real time.
- Choose-your-own-adventure story engine — 25 scenes, 4 endings, branching narrative powered by AI — built as an interactive web experience.
- 50+ pages of theological content generated in a single working session — not garbage output, but structured, referenced, publication-ready material.
Infrastructure
- 2 helpdesk instances deployed on production servers — configured, customized, and live.
- Interactive map with 70+ GPS-mapped data points and 4 switchable layers — a custom web application for visualizing geographic infrastructure.
- Full security audit across 5 web properties with credential rotation — identified vulnerabilities, fixed configurations, rotated exposed credentials.
- Custom MCP integrations connecting AI directly to business databases — no more copy-paste workflows.
- Client onboarding kit — a 4-document professional package, customized per client, delivered and in use.
The Math
A year ago, producing this output would have required:
| Role | Typical Rate | Hours Estimated |
|---|---|---|
| Web developer | $80-120/hr | 200+ |
| Telecom specialist | $100+/hr | 40+ |
| Content team (2-3 writers) | $50-80/hr each | 300+ |
| Game developer | $90+/hr | 100+ |
| Security consultant | $150/hr | 20+ |
| GIS/mapping developer | $90/hr | 30+ |
| Voice/audio engineer | $75/hr | 20+ |
| Project manager | $80/hr | 80+ |
Conservative total: 800+ hours of specialist labor across 6-8 professionals.
What I spent: $300/month in AI subscriptions. A desktop computer I already owned. Three months of evenings and weekends.
What This Actually Proves
I’m not listing this to brag. I’m listing it because the standard objection to AI adoption — “it’s not ready for real work” — is no longer defensible.
This isn’t demo output. These aren’t prototypes sitting in a folder. Every item on this list is live, deployed, and being used by real people.
The tools that made this possible:
– Claude Pro/Max ($20-200/month depending on usage)
– Claude Code (included with subscription — terminal-based AI that reads codebases and executes)
– Ollama ($0 — local AI for privacy-sensitive work)
– Standard web hosting (~$15/month)
That’s it. No proprietary platforms. No enterprise contracts. No venture capital.
The Gap Is Now Visible
In January, the difference between companies using AI and companies not using AI was theoretical. You could argue about it.
In April, it’s visible in output.
A one-person operation produced more deployable work in Q1 2026 than many 5-person teams produce in a year. Not because I’m better than those teams. Because the tools multiplied my output by an order of magnitude.
That multiplication is available to anyone. Right now. For the cost of a nice dinner per month.
The question every business owner needs to answer: if your competitor has access to the same tools and decides to use them — what does that do to your market position?
The answer isn’t “nothing.” And the tools are only getting better.
What’s Next
Q2 is going to be bigger. The tools just got another round of improvements. Claude can now dispatch teams of sub-agents. The context window holds a million tokens — the entire Bible, twice, with room for notes.
I’ll keep publishing what I build, how I build it, and what it costs. No gatekeeping.
If you want help applying these tools to your business, reach out. The first conversation is always free.
Matt Stoltz is the founder of Flower Insider Technologies, an AI-assisted managed IT company in southern Minnesota. He builds with AI every day and shares everything he learns.