Every organisation I work with has at least one. The AI champion. The person who got excited early, ran a few experiments, demonstrated value, and then got informally anointed as the go-to person for all things AI. They didn’t ask for the role. It just happened.

Six months later, that person is exhausted, resentful, and considering a job change.

I’ve watched this pattern repeat across dozens of organisations. The AI champion model sounds logical — find your enthusiasts, give them room to lead, and let innovation bubble up. But without proper support structures, you’re not creating a champion. You’re creating a single point of failure who’s doing two jobs for the salary of one.

The three ways AI champions fail

They drown in requests. Once word gets out that someone “knows AI,” every department comes knocking. Can you look at our process? Can you evaluate this vendor? Can you show my team how to use ChatGPT? Can you build us a prototype? The champion’s actual job — the one they were hired and paid to do — gets squeezed into evenings and weekends.

They lack authority. Champions can advocate, but they usually can’t approve budgets, allocate headcount, or mandate process changes. They end up in an exhausting cycle of persuasion — convincing individual managers to try things, one conversation at a time, without any organisational weight behind them.

They hit the expertise ceiling. An enthusiastic operations manager who’s great with prompting and understands business processes isn’t a machine learning engineer. At some point, the questions get technical enough that the champion is out of their depth. Without access to technical resources, they either wing it (dangerous) or become a bottleneck (frustrating).

The structural problem

The fundamental issue is that most organisations treat AI championing as a personality trait rather than an organisational function. They identify someone with passion and aptitude, say “go forth and innovate,” and assume that enthusiasm will carry the initiative.

It won’t. Enthusiasm is a starting fuel, not an engine. And individual passion can’t compensate for structural gaps in resourcing, authority, and technical support.

I spoke with specialists in this space recently who described the same pattern across their client base — organisations that appoint AI champions without budgets, without dedicated time allocation, and without clear reporting lines consistently see those champions disengage within 6-12 months. The problem isn’t the people. It’s the system around them.

What proper support looks like

If you’re serious about internal AI championing — and you should be, because grassroots adoption is how AI actually scales — here’s what the support structure needs to include.

Dedicated time allocation

This is non-negotiable. An AI champion needs a minimum of 20% of their working hours formally allocated to the role. Not “squeeze it in when you can.” Not “we’ll backfill your other work” (you won’t). A written agreement, endorsed by their manager, that one day a week is for AI activities.

Some organisations go further and make it a 50% or full-time role. That’s even better if the workload justifies it. The point is that championing AI is real work that takes real time, and pretending otherwise is dishonest.

A budget, even a small one

Champions need the ability to spend money without submitting a business case for every $500 tool subscription. A discretionary budget of $5,000-15,000 per year covers pilot tool licenses, training courses, conference attendance, and small-scale experiments. It’s a rounding error in most corporate budgets but transforms what a champion can actually do.

Technical mentoring

Pair your champion with someone technical — an internal data scientist, an external consultant, or even a vendor contact — who can answer the questions that go beyond the champion’s expertise. A monthly one-hour session where the champion can bring their toughest challenges is often sufficient.

This isn’t about making the champion technical. It’s about giving them a safety net so they don’t over-promise on things they don’t fully understand.

Executive sponsorship

Champions need someone with authority in their corner. An executive sponsor who can clear roadblocks, approve cross-departmental collaboration, and — critically — signal to the broader organisation that AI championing is valued and supported.

The sponsor doesn’t need to be deeply involved. Monthly check-ins and a willingness to make phone calls when things get stuck is usually enough. But the champion must know they have top-cover.

A community of practice

One champion in isolation is fragile. Three or four champions connected in a community of practice is resilient. They share learnings, divide problems, provide moral support during setbacks, and collectively build momentum that no individual could sustain alone.

The community doesn’t need to be formal. A Slack channel and a fortnightly lunch meeting is enough. What matters is that champions aren’t operating alone.

The role of L&D

This is where learning and development teams should be stepping up — and mostly aren’t.

L&D has a natural role in supporting AI champions. Curating training resources, facilitating community meetings, tracking skill development, connecting champions across business units, and measuring the impact of champion-led initiatives. These are core L&D capabilities applied to a new context.

The Australian HR Institute has published guidance on building internal capability networks that translates directly to the AI champion model. The principles of peer learning, structured development, and organisational support aren’t new. What’s new is the application domain.

Yet in most organisations, L&D isn’t involved in AI championing at all. It sits with IT or innovation teams, disconnected from the people development infrastructure that could sustain it.

Signs your champion programme is failing

Watch for these signals:

• Your champion has stopped running experiments and is only answering other people’s questions
• They’re working evenings and weekends on AI stuff because their day job hasn’t been adjusted
• They’ve started saying “I’m not sure that’s my role” when asked about AI topics
• Other departments have started hiring their own AI contractors because the champion is a bottleneck
• The champion’s manager is complaining about their “real work” slipping

If you’re seeing two or more of these, your support structure needs urgent attention.

Getting it right

The organisations I’ve seen succeed with AI champions treat the role as a defined, supported function — not an informal hobby. They allocate time, provide budgets, offer technical mentoring, assign executive sponsors, and connect champions into communities.

None of this is complicated. None of it is expensive. It’s just intentional.

The alternative — burning through your most enthusiastic people and wondering why AI adoption stalls — is far more costly. Your champions are willing. Give them what they need to succeed.