Lessons · 13 October 2026 · 2 min read
Why We Stopped Building MingoolAI, and What It Taught Us
MingoolAI started as a fix for a problem we had ourselves: an inbox and a calendar that took more of the day than the work did.
We built an AI Chief of Staff for solo founders. It connected to Gmail and calendar, processed everything overnight, and sent a short briefing each morning covering what was urgent, what could wait, and what needed a decision. Amy, a Chrome extension, carried that context through the browser during the day.
We soft-launched in April 2026 and had a first paying user on launch day. We designed three pricing tiers, set up a small paid beta programme, and planned our distribution. Then we retired it. Here is why.
What worked
- The problem was real. Someone paid on launch day, which told us the pain was worth solving.
- We shipped the whole loop: overnight processing, a daily briefing, and a browser extension that kept context during the day.
What went wrong
- Our growth depended on a gate we did not control. Apps that read Gmail have to pass Google's security review (CASA Tier 2). Until that cleared, it limited how many new users we could bring in, so every growth plan waited on a process outside our hands.
- A product needs an owner whose full-time job is that product. MingoolAI ran alongside a busy agency. When we made the decision, we were still looking for a dedicated head of product.
- Splitting focus costs more than it looks. We were running client work, a new product and other initiatives at once. Each got part of our attention, and none got enough.
What we would do differently
- Start the platform review first. If your product depends on a permission or a review from a platform, begin that process before you build, and plan around its timeline.
- Do not start a second product until the first one has an owner.
- Set a clear test for continuing, with a date, before you launch.
Where we are now
We have retired MingoolAI to focus on two things: building software for clients, and Resolve, our customer support intelligence product.
MingoolAI stays in our portfolio with its limits written beside it. A record of what did not work is useful to us and, we hope, to you.