Something interesting is happening in corporate L&D teams, and I don’t think enough people are talking about it. A new role is emerging — part technical advisor, part change manager, part therapist — and organisations are calling it the “AI Coach.”

It’s not what you’d expect. And it’s solving a problem that traditional training can’t.

The problem that created the role

Here’s the situation most organisations are in right now. They’ve deployed AI tools — maybe copilots for writing, maybe analytics dashboards with AI features, maybe industry-specific applications. The tools are live. The training has been delivered. And… adoption is patchy at best.

Some people are using the tools daily and getting real value. Others tried them once, got a weird result, and went back to their old way of doing things. A third group never really engaged at all.

The traditional L&D response would be more training. Another workshop. A refresher course. An e-learning module with a quiz at the end.

But that’s not working, because the barrier isn’t knowledge. People know the tools exist. They’ve been shown how to use them. The barrier is confidence, habit, and context. And those require a fundamentally different intervention.

What an AI Coach actually does

The AI Coaches I’ve been observing sit within L&D teams but operate more like embedded consultants. They work with individuals and small teams, usually in sessions of 30-60 minutes, addressing specific work challenges.

Here’s what a typical week might look like:

Monday. A marketing manager is struggling to get useful outputs from the company’s AI content tools. The AI Coach sits with them, works through their actual brief, shows them how to refine their prompts, and helps them develop a workflow that fits their existing process.

Tuesday. A finance team is nervous about using AI for forecasting because they don’t trust the outputs. The AI Coach runs a session showing them how to validate AI-generated forecasts against historical data, building their confidence through evidence rather than reassurance.

Wednesday. A project manager has identified a potential AI use case but isn’t sure whether it’s feasible. The AI Coach helps them think through the data requirements, connects them with the right technical team, and helps them write a business case.

Thursday. A senior leader is about to present an AI strategy to the board and wants to pressure-test their thinking. The AI Coach plays devil’s advocate, asking the tough questions a board member would ask.

The common thread is that none of these interactions look like traditional training. They’re contextual, personalised, and embedded in real work. That’s what makes them effective.

Why this isn’t just another name for “AI Trainer”

I want to be clear about the distinction because I’ve seen some organisations rebrand their existing trainers as “AI Coaches” without changing anything. That misses the point entirely.

An AI Trainer delivers structured content to groups. They teach concepts and demonstrate tools. Their measure of success is whether people completed the training.

An AI Coach works one-on-one or in small groups, addressing specific challenges in context. They don’t follow a curriculum — they follow the learner’s needs. Their measure of success is whether people actually changed their behaviour.

The skillset is different too. A good AI Coach needs:

• Broad AI literacy. Not deep technical expertise, but a solid understanding of what different AI tools can and can’t do across multiple domains.
• Coaching skills. Actual coaching skills — active listening, asking good questions, helping people find their own solutions rather than prescribing answers.
• Business acumen. They need to understand the business context well enough to recognise when an AI application makes sense and when it doesn’t.
• Emotional intelligence. People have complicated feelings about AI. Fear of job loss, frustration with new tools, excitement about possibilities, scepticism about hype. A good AI Coach navigates all of that.

The evidence that it works

I’m aware of about a dozen organisations in Australia that have created formal AI Coach roles in the past twelve months. The results are preliminary but encouraging.

A Harvard Business Review piece from late 2025 highlighted that organisations with embedded AI support saw 2.3x higher tool adoption rates compared to those relying solely on group training. That aligns with what I’m seeing locally.

One financial services firm I’m connected with tracked AI tool usage before and after introducing two AI Coaches. In the six months prior, active daily users of their AI analytics platform plateaued at 34% of the intended user base. Six months after the coaches started, it was 71%.

More telling than the usage numbers were the qualitative results. In feedback surveys, the most common comment was some variation of “I didn’t know I could use it for that.” The coaches were helping people see applications that the original training never covered, because those applications were specific to individual roles and workflows.

How to set up the role

If you’re considering adding AI Coaches to your L&D team, here’s what I’d suggest:

Start with one. Don’t hire a team. Start with one person, embed them in the business area with the lowest AI adoption, and learn from the experience.

Recruit internally first. The best AI Coaches I’ve seen came from within the organisation. They understand the culture, the processes, and the politics. You can teach AI skills to a good internal coach faster than you can teach organisational knowledge to an external AI expert.

Give them a dotted line to IT. The AI Coach needs a strong relationship with the technical team. They’ll frequently need to check whether something is feasible, flag bugs, or relay user feedback. A formal connection to IT or the data team makes this easier.

Measure behaviour change, not satisfaction scores. The happy sheet after a training session is meaningless. Track tool adoption, track productivity metrics, track whether people are coming back for follow-up sessions. Those tell you whether the coaching is working.

Protect their time. It’s tempting to pull them into other L&D activities — running workshops, creating e-learning content, managing the LMS. Resist that temptation. Their value is in the one-on-one and small-group work. If they’re spending half their time on admin, you’ve wasted the investment.

Where this is heading

I think AI Coaching will become a standard L&D function within two years. As AI tools proliferate across every business function, the need for contextualised, ongoing support will only grow. Group training will always have its place for foundational knowledge, but the real adoption gains come from the personalised, embedded approach.

The organisations investing in this now will have a meaningful head start. Not just in tool adoption, but in building the kind of AI-fluent culture that makes future technology adoption faster and smoother.

It’s one of the more promising developments I’ve seen in corporate learning in a long time. And I don’t say that lightly.