Written by Michiel R. De Boer
We have made this mistake before
Look back at the big technology decisions of the last twenty-five years.
- The internet. We asked for an online presence. We got our brochure on a website.
- We moved to the cloud to save cost. We moved our servers as they were, and the savings in the business case did not always show up.
- We invested in faster delivery. We bought the tools, renamed a team and kept the same change board.
- Digital transformation. We funded a paperless organisation. The paper form became a PDF that still needed three signatures.
- We asked for use cases. Now chatbots sit on top of broken processes, and reports nobody reads get written faster.
The technology was real every time. What went wrong was the instruction. We pointed a powerful tool at the way we already worked, and it did exactly what we asked. It made the old way faster, bigger and harder to change.
When a directive loses its purpose
Most of these programmes started with a good intention at the top. A board sees a shift in the market and wants the organisation to move. So it sets a target: every department must show an AI use case this quarter. Or half of all processes automated by year end.
Then the target travels down. Each layer keeps the deadline and loses a bit of the reason. By the time it reaches the people doing the work, the question is no longer why but how many. Nobody asks whether a process should still exist. They ask how quickly they can report it as automated.
That is not a failure of the teams. They did what was asked. The directive described an activity, not an outcome, and it left no room for anyone to say: this should be removed, not automated.
Mirror question: look at the last technology target you signed off. Did it describe a result for the customer or the business, or only an activity to report?
Asking for the tool instead of the result
Real innovation starts with a problem worth solving. Board-driven innovation often starts with a tool that needs a use case.
You can hear it in the workshop titles. “Where can we use AI?” finds places for the tool. “What slows our customers down?” finds value. The first is easier to report upward. The second pays back.
That is how the battery ends up on the axe. Nobody asked the person swinging it what was hard about the job.
There is a second trap. Once a pilot is linked to a board target, nobody dares to stop it, because stopping looks like failure. So weak ideas do not fail small and cheap. They get scaled.
Automating the weight we added ourselves
Transformation should change how work is done. Most automation programmes only change where it is done. The same seven approval steps move into a workflow tool. The same handoffs between teams move from email to tickets.
That is the wooden axe head. Every part is there, just in the wrong place.
It is worth being honest about where that weight came from. Many of those controls were added at senior level, after an audit finding, an incident or a question from the regulator. Each one made sense at the time. Together they slowed the organisation down. Automating them does not remove the weight. It locks it in, at a higher cost.
Simplify first. Remove what no longer adds value. Then automate what is left. It is usually less than the target, and worth much more.
Reporting what is easy to count
Look at the last few board packs on digital or AI. Number of bots. Number of use cases. Licences rolled out. Percentage of processes automated.
These numbers grow easily and look good on a slide. They say nothing about whether customers are served better, costs are lower or people now spend their time on work that matters. They measure the battery, not the cut.
The harder questions are the useful ones. Did lead time go down? Did errors go down? Did people get time back, and what did they do with it? If management cannot answer these, the automation may be adding cost instead of removing it. And those answers should shape the next directive, not only the next target.
What the top table sets in motion
None of this is really about technology. It is about four human areas that boards and senior management shape every day, often without noticing.
- Behaviour. People do what gets noticed and rewarded. Reward “automated” and you get automation. Reward better outcomes and you get better questions.
- Collaboration. Processes cross team borders. Budgets usually do not. Each unit automates its own piece and the gaps between them stay. In hierarchical cultures, people also rarely feel safe telling the top that a directive does not make sense.
- Leadership. Senior leaders steer through the questions they ask. Ask “how many?” and you get numbers. Ask “what changed for the customer?” and you get value. And protect the person who dares to say stop.
- Motivation. A directive buys compliance. A purpose earns commitment. Compliance delivers the wooden axe head on time. Commitment is what notices it will not cut.
Questions for your next board meeting
- What problem are we solving, and for whom?
- Should this process exist at all, before we automate it?
- Which of our own controls are we about to lock in?
- How will we know it worked, beyond “it is automated”?
- Who in the organisation can tell us to stop, and would we listen?
Before the next swing
AI may be the most powerful tool boards have been asked to steer so far. That makes aim more important, not less. A strong tool pointed at the wrong target does more damage, faster, and it is harder to undo.
