Direction · 3 min read
AI leadership
Leading on AI is not a procurement decision. It is choosing a direction, owning the data, building the people and being willing to be measured. Here is what it asks of a leader.
B57 Group
Every government now says it is serious about AI. Most of them mean they have bought something. A platform, a pilot, a partnership announced with a photograph. Leadership is something else, and it is rarer, because it is harder.
AI leadership is deciding where the country is going with this technology, taking ownership of the things that cannot be bought, and accepting that the results will be visible. Here is how to tell whether it is happening.
Three tests
Direction. Can the leader say, in two sentences, what AI is for in their ministry? Not a list of use cases. A purpose. “We will cut the time to register a business from thirty days to three.” “Every district health officer will have a decision-support tool in their own language.” If the answer is a vendor’s slide, there is no direction yet.
Ownership. Who holds the data? Where do the models run? Who can switch them off? If the honest answer is a company in another jurisdiction, the leader has adopted AI but does not own it. Ownership is the test that separates sovereign capability from a subscription.
Measurement. What number will move, and when will it be published? A leader who is serious sets a target that can be missed in public. A leader who is not serious talks about transformation.
A leader who is serious sets a target that can be missed in public.
The first hundred days
Direction first. Choose one outcome that citizens will notice and that the ministry can control. Write it down. Say it out loud.
Then inventory what you hold. The data, the systems, the people who understand both. Most ministries discover they have more of the first and less of the third than they thought. That gap is the real constraint, and it is fixed by training, not purchasing.
Then build one thing, small, with a team that includes your own officers. Not a strategy. A working system in one office that moves the number you chose. Visit it. Use it. When it works, say so, and when it does not, say that too. The credibility of everything that follows rests on the honesty of the first result.
Leadership at every level
The mistake is to think AI leadership lives only at the top. A director who redesigns one process is leading. A district officer who learns the tools and teaches her team is leading. A clerk who flags that the model got a case wrong, and is listened to, is part of a leadership culture.
This is why we put so much weight on training at scale. A government with a handful of AI experts and thousands of people who have never touched the tools is not led. It is dependent. A government where practical AI skills are ordinary, where the people doing the work can shape the systems that do it with them, has leadership distributed through the whole organisation. That is resilient in a way no vendor relationship can be.
The line
The bar under our name marks the year this country decided to lead rather than be led. The decision in front of today’s leaders is smaller in scale and the same in kind. Intelligence is infrastructure. The governments that build it will set the terms for the ones that rent it.
Lead on AI the way the country has led before: choose the direction, own the ground, build the people, and be willing to be measured on the result.