← Lorenzo Ventures.
Case study · MODA
lorenzo@lorenzoventures.co
Case studyBuilding Moda with an AI agent teamFeb — Sep 2026

One founder. Eight agents.An app on Apple's doorstep.

How Moda went from chat-assisted prototype to App Store submission — and what changed when we stopped prompting and started delegating.

February 2026
First commit
May 2026
Agent team replaces chat workflow
September 2026
Submitted · in App Review

Phase one

The chat era

Moda began the way most AI-assisted projects begin now: a founder, a day job, ChatGPT for thinking, and Cursor for writing code. First commit, February 2026.

It worked — until it didn't. Chat tools are powerful but stateless. Every session started from zero: re-explaining the codebase, re-establishing conventions, re-discovering decisions already made. The founder was the memory, the router, and the bottleneck. Progress moved exactly as fast as one person's free evenings.

Phase two

The agent team

In May 2026, we replaced the chat workflow with a team of autonomous AI agents, each with a name, a role, and a schedule:

  • Product managerMo

    Reviews every pull request, runs standups, keeps the roadmap honest.

  • Frontend & backendStitch & Loom

    Frontend and backend engineers.

  • VisionPix

    Vision and photo-intake pipeline.

  • Release & infraAtelier

    Release engineering and infrastructure.

  • IntelligenceMuse

    Intelligence.

  • EngagementBell

    Engagement.

  • MonetizationCoin

    Monetization.

  • HumanFounder

    Sets direction, makes product calls, approves releases.

The agents run around the clock on their own schedules. They claim tasks from a shared board, work in isolated git worktrees, open pull requests, and pass a full CI gate — typecheck, lint, and 1,900+ automated tests — before the PM agent will merge. A circuit-breaker system halts the line when anything looks wrong.

The founder sets direction, makes product calls, and approves releases. The team does the rest — including while he sleeps.

What shipped

All numbers pulled from the repo and fleet logs, current as of September 2026 — nothing estimated.

848

commits, 840+ merged pull requests — the overwhelming majority authored, reviewed, and merged by agents

550+

tracked tasks executed across design, engineering, QA, release, and App Store operations

1,919

tests across 220 test suites — written and maintained by the same agents that write the features

9

PRs merged in a single day at peak — while the founder was at work

Feb Sep

idea to App Store submission, nights-and-weekends founder time only

Outcome

What actually changed

The shift wasn't speed alone — it was continuity. Agents hold context across sessions, so decisions stick.

Quality became systematic instead of heroic: the PM agent has rejected CI-green pull requests for visual defects it caught by measuring font metrics against screen widths — review depth a tired human skips. And the founder's job changed from typing code to making decisions, which is the only part that ever required him.

Chat tools made one person faster. Agents made one person a team.

1,919 tests. 840+ pull requests. One human.