The situation
iMaintain knew exactly where engineers lose time and how much knowledge lives in people’s heads instead of systems. What they didn’t have was a technical team to turn that into software, without losing what made the idea special. They were experts in maintenance, not in code.
What we questioned first
- What should the AI do?Rather than rushing to build, we ran a short discovery sprint to map the conversations engineers would have with an assistant and the data it would need. The output was an AI use case document, a data blueprint and an architecture plan.
- Does it work before we commit?Two weeks to build the first working iMaintain Brain against simulated maintenance data. Something the founders could touch, test, break and believe in before money went into the full product.
- How do we keep the founders in the loop?The MVP was built sprint by sprint with the founders testing features as they landed, so it never felt like two separate teams.
What we built
A production MVP of an AI-powered maintenance management system: intelligent work order management, with the iMaintain Brain integrated throughout to help engineers troubleshoot, find manuals and pull up maintenance history instantly.
What changed
Senior engineers’ knowledge no longer has to stay in one person’s head: it can be captured, shared and used across a whole team. The founders went from an idea to a working product in six months.
Fluff have been brilliant to work with, a proper partnership rather than just a dev team. They’ve taken the time to understand what we’re building, challenged ideas in the right way, and delivered solid, scalable code. Communication’s been spot on and they genuinely care about the product. Would definitely recommend them.
How we’d work on this today. For a founder without a CTO, a fractional product team challenges the idea, plans it properly and builds it in small, tested releases. Today that starts with a Discovery Sprint and carries on through the queue. Start with a Discovery Sprint
