Most of what gets written about AI in leadership development is about spotting future leaders, scoring competencies or predicting who will succeed. That is useful for an HR team. It is not much use if you run leadership programmes for clients and want to know what AI would actually do inside one of them.
I build programme portals and AI coaches for consultancies that run leadership, sales and wellbeing programmes. So this is the view from inside the programme: what AI does well there, what it should never do, and what it takes to add it to a programme you already run.
The short answer
Inside a leadership programme, AI earns its place in the weeks between workshops. It does five jobs well:
- Answering delegates' questions from the programme material, at the moment they need it.
- Letting delegates practise a hard conversation before they have the real one.
- Sending nudges based on what each delegate has actually done, not on the calendar.
- Turning reflection into material for the next session, so the facilitator walks in knowing what the group is stuck on.
- Showing the client what changed, with confidence and progress data they can read in a minute.
It should not replace the facilitator, give advice from outside your model, or deal with anything personal on its own. Those stay with people.
Where leadership programmes lose their value
Anyone who has run a leadership programme knows the pattern. The workshop goes well. People leave with a model, a workbook and good intentions. Then they go back to a full inbox.
Three weeks later, one of them is about to give difficult feedback to someone in their team. That is the moment the programme was designed for. But the workshop feels a long way off, the workbook is in a drawer, and the facilitator is with another client. So they do what they have always done.
This is the gap AI fits into. Not the workshop, which works because people are in the room together. The space between workshops, where the programme either turns into new habits or quietly fades. There is a full follow up plan for those weeks in how to reinforce leadership training.
Five things AI does inside a leadership programme
1. A coach beside every session
The most useful single piece is an AI coach trained on the programme's own material: the session videos, the workbook, the slides, the models your facilitators teach. A delegate who is stuck can ask it a question at eleven at night and get an answer in the programme's language, with a pointer to the session and page it came from.
Many of these questions are simple. Where is the workbook for session three? What were the four steps of that feedback model again? Others are applied: how do I use this with someone who gets defensive? A coach built on the programme answers both, and it answers them the way the programme teaches, not the way a general chatbot would.
2. Practice before the real conversation
Leadership is learned by doing, and the hardest things to practise are conversations. A delegate can rehearse the feedback conversation, the difficult one-to-one or the opening of a board meeting with the AI playing the other person, then get feedback based on the programme's model.
It is not the same as practising with a skilled facilitator or a peer. It is a lot better than going in cold, and it is there the evening before, which is when people actually prepare.
3. Nudges triggered by progress
Most programme emails go out on a schedule: week two, week three, week four. They land whether a delegate is flying or has not logged in for a fortnight.
Nudges work much better when they react to what someone has done. Three days without activity gets a short, warm message with the one thing worth doing this week. Finishing the first part of the programme gets a note that marks it. Every message is written in your voice and approved by you before anything goes out.
4. Reflection that feeds the next session
Reflection questions in a workbook are easy to skip. A short prompt after each session, which a delegate can answer in a few lines on their phone, is not.
With the delegates' consent, the facilitator can see the themes across a cohort before the next workshop: what people tried, where they got stuck, which part of the model nobody is using yet. The session starts from where the group actually is.
5. Evaluation the client can read
The question every client eventually asks is "did it work?" AI makes the answer easier to collect: confidence ratings at the start and end, session ratings, which sessions people come back to, which questions come up most. The client gets a short picture of their cohort instead of a pile of feedback forms.
What AI should not do in leadership development
The same tools can do real damage to a programme if they are set up carelessly. These are the four lines I hold on every build.
It should not replace the room. A leadership programme works because people practise with each other, hear each other's situations and are challenged by someone experienced. AI supports that. It does not stand in for it.
It should not answer from outside your model. A general AI tool will happily answer a leadership question with advice from anywhere on the internet. If your programme teaches one way of giving feedback, a coach that suggests a different one undoes your work. A coach built for a programme answers from that programme's material and says so when a question goes beyond it.
It should not handle personal situations alone. When a delegate brings something personal, sensitive or high stakes, the right answer is a person. A well built coach recognises those moments and passes them to the facilitator, with context, instead of guessing.
It should not become a surveillance tool. If delegates think what they ask the coach will be reported to their manager, they stop asking real questions. Agree up front what the client sees (usually totals and themes across the cohort) and what stays private, and tell delegates in plain words.
Why not just let delegates use a general AI tool?
Many of your delegates already use a general AI chatbot at work, so this is usually the first thing consultancies ask about AI in leadership training. The difference is not cleverness. It is three things.
It speaks your programme. It uses your models, your terms and your examples, so every answer reinforces what the delegate learned instead of competing with it.
It knows its limits. It stays inside the programme and passes the rest to a person.
It is yours. For a consultancy, the programme is the product. A coach built on it, in your brand, becomes part of what you sell and renew, rather than a free tool your client can use without you. And your material stays under your control: it is not used to train public AI models. What "in your brand" should cover is in the white-label platform guide.
What using AI in leadership development takes
Less than most people expect, because the material usually exists already.
Gather what you have. Session recordings, workbooks, slides, facilitator notes, the questions delegates ask most. Nothing needs to be rewritten first.
Decide the boundaries. What the coach should answer, what it should always pass to a facilitator, how it should sound, and what the client gets to see.
Start with one cohort. Run it with a single group, read what delegates ask, and adjust before you roll it out to every client.
Measure from day one. Take a confidence rating at the start, so there is something to compare against at the end.
On the programmes we build, the full system of portal, AI coach and nudges usually goes from your material to launch in about six weeks. If the programme is still classroom only, taking a leadership programme online covers what to move and what to keep live, and what a programme portal costs puts numbers on the options.
Questions consultancies ask
Will delegates actually use it? They use it when it is one tap away and answers properly the first time. That is why it belongs beside each session in the programme itself, not in a separate tool they have to remember.
Does it replace our facilitators? No. It takes the repeat questions and the evening-before preparation off their plate, so their time goes to the work that needs them.
Can it carry our brand? Yes. The portal, the coach and every message can carry your name, your colours and your voice. Delegates experience it as your programme.
What happens to delegates' data? You decide what is collected and who sees it, and the setup should follow GDPR from the start. Ask any provider where the data is stored, how long it is kept and whether it is used to train AI models. The answers should be short and clear.
Is it worth it for a small programme? It pays off most where the same programme runs for several cohorts or several clients, because the material is built in once and used again every time.
Where to start
If you run a leadership programme, the useful question is not "should we use AI?" It is "what happens to our delegates in the three weeks between workshops?" AI is the most practical answer to that question I know, as long as it is built on your programme and knows where its job ends.
To see the whole system together, the programme portal page shows the portal, the AI coach and the nudges as one. For coaches running their own programmes, the AI coach is the same idea on a smaller scale.
If you'd like to see what this looks like with your own material, I'll build a companion from your website, free. See your own companion
Robert, founder of Thinkable Systems