Two people, the output of a much larger team
YoBo
A full live-event ticketing platform: direct ticket sales, reserved seating on interactive seat maps, seat upgrades, gifting, and voucher redemption for shows, tours, and attractions — across nine partner booking and ticketing systems, with real payments and real bookings every day. Built and run by two people: a developer and a product owner who had never written code before.
The AI angle
AI agents write the code, the tests, the reviews, and the release packages; people decide what gets built and approve every release. The result is roughly a thousand hours of equivalent engineering work a week, from a team of two.
- Direct tickets, reserved seating & seat maps, upgrades, gifting, vouchers
- 464 PRs and ~280,000 lines merged in the last 10 weeks
- ~1,000 hours/week of equivalent engineering output, from two people
- Builder, reviewer, and fixer agents — no agent approves its own work
- Test-first on money and bookings: 400+ backend, 240+ browser tests
- An agent monitors production after every release
- AI voice agent guides callers through redemption in real time
YoBo is the clearest picture of how Hansen House works day to day. It's a live ticketing platform with real money moving through it — and it's built and run by two people and a pipeline of AI agents, shipping at the pace of a much larger team.
The Platform
YoBo is a full live-event ticketing platform. Customers buy tickets direct, pick their seats on interactive seat maps, upgrade seats, send tickets as gifts, and redeem vouchers and pre-paid tickets — for shows, tours, attractions, and events. Under the hood it connects nine partner booking and ticketing systems, each with its own data model and quirks, behind one flow. It handles real payments and real bookings every day, and has been in production since 2020.
One feature among hundreds: an AI voice agent that takes inbound calls, triages the request, texts the caller a check-in link, and stays on the line narrating each step of the redemption until the booking is complete.
Two People, the Output of a Much Larger Team
YoBo is built and run by two people: a developer, and a product owner who had never written code before. Together they deliver roughly a thousand hours of equivalent engineering work every week — our estimate of what a traditional team would need to produce the same output.
In the last ten weeks (as of October 2026) they merged 464 pull requests and about 280,000 lines of code into the production codebase.
What Makes It Work
AI agents write the code, the tests, the reviews, and the release packages. People decide what gets built and approve every release. Five practices hold it together:
- Rules live in the repo. Every agent reads the same instructions before it touches anything. They require reading the documentation first, never guessing how a partner's system works, and using structured, classified errors. A short context guide sends agents to the right deeper document.
- The pipeline is a set of repeatable skills. Planning, coding, pull requests, review, merging, release gates, deploys, and production monitoring each run as a defined workflow. An urgent fix follows its own path, so it never ships unreleased work by accident.
- Builder and reviewer are separate. One agent writes the change. A different, read-only agent reviews it. A third makes only the fixes the review calls for. No agent approves its own work.
- Money and bookings are tested first. Payments, barcodes, partner connections, and check-in logic need a failing test before any code is written. The suite holds over 400 backend tests and over 240 browser tests, and the rules warn against tests that pass without checking anything.
- Production is watched by an agent too. After each release, a monitor reads partner logs, payment failures, and server health. A new kind of failure isn't considered handled until the monitor can recognize it.
What This Means for You
The product owner's contribution is knowing the business: the rules, the partners, and the customers. The agents turn that knowledge into working, tested code. The safeguards turn it into code that's safe to run in production. That combination is what we build for clients.