Where do you start when you do not know where to start?

When the leadership wants to move on AI without knowing where to start, the temptation is to choose a tool, or to launch the most visible project. You start with the solution before looking at the problem.

Yet the best first uses are rarely visible from the boardroom. They lie where information waits, gets lost or is keyed in again: between two departments, in an approval that holds everything up, in a task nobody likes doing but everyone depends on. To see them, you first have to represent how work really flows.

• How things really work differs from the organisation chart.The official channels are doubled by shortcuts, shared files and arrangements that only the teams know about.
• Bottlenecks often sit at the boundaries between departments.That is where information gets lost, waits or changes.
• Once the map is drawn, the priorities become clear to all.Everyone sees the same thing, and the debate is about choices rather than findings.

1. 01Choose a flow that matters, from order to invoice for example, and have it described by the people who keep it running.
2. 02At each step, note what waits, what is keyed in again and what depends on a single person.
3. 03Do not choose any tool before doing this exercise.

Most of our assignments begin with this situation: drawing the map before choosing anything.

• How do we choose our first use cases?→
• We already have ChatGPT or Copilot licences. Is that an AI strategy?→
• Are we behind?→
• Is our data ready for AI?→