Most businesses don't fail with AI because the technology isn't ready. They stall because they start in the wrong place — with a tool instead of a problem. Here is the checklist we use in our first conversation with every client.
1. List the work that repeats
Ask each team to write down tasks they do every week that follow roughly the same steps: answering the same questions, copying data between systems, preparing the same report. Repetition is where AI pays off first.
2. Score each task on three things
| Question | Why it matters |
|---|---|
| How many hours a month does it take? | The bigger the number, the bigger the saving. |
| How costly is a mistake? | Start where errors are easy to catch and fix. |
| Is the information already digital? | Paper and scattered spreadsheets add setup time. |
Pick the task with high hours, low risk and digital inputs. That is your first project.
3. Define "done" before you build
Agree on one or two measurable outcomes — for example "replies drafted for 80% of routine enquiries" or "weekly report ready by Monday 9 am without manual work". Clear targets make it easy to decide whether to scale up.
4. Prototype on real work, with a person reviewing
Run a small version on genuine tasks for two to four weeks. Keep a human approval step before anything is sent to a customer or written to a system. You'll learn more from a week of real use than from a month of planning.
5. Roll out, then keep improving
Once the prototype meets its targets, add access controls, monitoring and short training for the people who use it. Review it monthly: models, prices and your own processes change, and small adjustments keep costs low and quality high.
Start with one task, prove it, then repeat. The second project is always faster than the first.
If you'd like a second pair of eyes on your list, book a free discovery call.