Every week, someone asks us to "implement AI" without being able to describe what problem they're trying to solve.
The tool-first trap
It usually starts with a demo. Someone on the team sees a flashy product, brings it to the next meeting, and suddenly there's a mandate: "We need this." No diagnosis, no requirements, no understanding of what it would replace or whether anyone would use it.
We've seen this play out dozens of times. The pattern is always the same:
- A tool gets purchased or built
- It doesn't fit the actual workflow
- It gets abandoned within months
- Everyone concludes "that technology doesn't work"
The technology was fine. The process was backwards.
Start with the pain
Before we talk about tools, we ask three questions:
- What's the bottleneck? Where does work slow down, get stuck, or fall through the cracks?
- What's the cost? Not just money — time, attention, error rates, missed opportunities.
- What does better look like? Not "AI-powered" or "automated" — what specific outcome changes?
Only after we can answer these clearly do we start thinking about solutions. Sometimes the answer is AI. Sometimes it's a spreadsheet. Sometimes it's a conversation that should have happened six months ago.
The right tool for the job
We're not attached to any particular technology. We use AI when it's the best fit. We use automation when it makes sense. We use low-code tools when speed matters more than customisation.
The goal isn't to use the most impressive technology. The goal is to solve the problem in the simplest way that actually works.
What matters is whether the solution fits the problem, whether the team can maintain it, and whether it delivers real value — not whether it looks good in a pitch deck.