9 March 2026
Why Every Board Deck Needs an AI Exposure Section
Board decks need a standing AI exposure section, sitting right next to the financials, and I don’t think that’s optional anymore for any company with real AI dependency.
Key takeaways
- Model dependency and vendor lock-in are governance risks, not engineering details
- Boards that can’t interrogate AI cost base aren’t doing proper oversight anymore
- An AI exposure slide should be as routine as cash runway at every board meeting
- The slide only needs four data points to be genuinely useful to a board
- Adding this now is far cheaper than explaining an AI cost shock after the fact
I’ve sat on both sides of the board table, first as a founder presenting to my own board at Obby and Baluu through to exit, and now as an advisor sitting on and supporting boards through VAQA. In both seats, I’ve watched AI dependency grow inside a business without ever earning a permanent line in the board pack. This piece sets out exactly what I think that slide should contain and why leaving it out is now a governance gap.
Why AI exposure belongs in board governance
Boards exist to interrogate the risks a company is carrying, and AI dependency is now one of the largest operational risks most software companies hold. Cash runway gets a slide every meeting because running out of cash can end a company quickly. Model dependency deserves the same standing, because a pricing change, an outage, or a policy shift from a single AI provider can hit a product just as hard.
I don’t think this is a controversial claim once you say it out loud. The gap exists mostly because AI moved faster than board reporting templates did, not because directors don’t care about the risk.
What most board decks are missing right now
In my experience reviewing decks as part of board advisory work, AI usually appears in one of two places: a product roadmap slide about new features, or a cost line buried inside general opex. Neither gives a board what it actually needs to do its job.
A director looking at those slides can’t answer basic questions. Which providers is the company dependent on. What happens to margin if that provider’s pricing changes. What’s the fallback if that provider has an outage. None of that is answerable from a roadmap slide or a lump sum cost figure.
What belongs on the AI exposure slide
I keep this deliberately short, because a slide nobody reads doesn’t improve governance. Four data points, updated every meeting, cover the risk properly:
| Data point | Why it matters to the board |
|---|---|
| Primary model providers in production | Shows concentration risk at a glance |
| AI cost as a percentage of total opex | Tracks trend, not just absolute spend |
| Fallback or multi provider status per critical use case | Shows whether an outage would stop the product |
| Material pricing or policy changes since last meeting | Flags anything the board needs to act on now |
That table takes a few minutes to prepare once the underlying tracking exists, and it turns a vague sense of “we use a lot of AI” into something a board can actually govern against.
Who should own the AI exposure slide
Ownership matters as much as content. I’d put this with whoever already owns the cash runway and financial slides, working directly with whoever runs engineering, rather than leaving it as a standalone technical update from the CTO.
That pairing matters because AI exposure is a financial risk as much as a technical one, and splitting it across two disconnected slides is exactly how it falls through the cracks. One owner, one slide, reviewed every meeting, is what makes it stick.
Boards that ask for this consistently tend to get better answers over time, because the company starts tracking it properly between meetings rather than scrambling to answer it once a year.
Why this can’t wait for a crisis
The cost of adding this slide now is close to zero: a short conversation with the finance and engineering leads, and a template that gets reused every meeting after that. The cost of not having it is a board that finds out about a material AI dependency only when something has already gone wrong.
I’ve advised companies through exactly that scenario, where a pricing change or an outage forced an urgent board conversation that could have been a routine agenda item months earlier. Every one of those situations was avoidable with a slide that took ten minutes to prepare.
Good governance is mostly about not being surprised by things you could have seen coming, and AI exposure is now squarely in that category for any company with real product dependency on it.
Put AI exposure on the agenda permanently
I don’t think AI exposure reporting is a nice addition to board packs. For any company with meaningful AI dependency, it’s a governance requirement that’s simply arrived a little later than it should have.
The format doesn’t need to be elaborate. Four consistent data points, owned jointly by finance and engineering, reviewed at every meeting, is enough to turn a blind spot into a managed risk.
Boards that build this habit now will be the ones asking sharp questions before a pricing shock hits, rather than explaining one after it already has.
Through my Board Advisory work at VAQA, I help boards build exactly this kind of standing governance practice around AI dependency. If your board pack doesn’t currently include an AI exposure section, get in touch and I’ll help you build one that actually gets used, and you can read more about what I advise on and about Tom.
Frequently Asked Questions
What exactly should an AI exposure slide include?
At minimum, primary model providers in production, AI cost as a percentage of total opex, fallback status for critical use cases, and any material pricing or policy changes since the last meeting. Those four points give a board enough to ask informed questions.
Who should present the AI exposure slide at a board meeting?
I’d recommend whoever already owns the financial slides, working closely with the engineering lead, rather than treating it as a separate technical update. That keeps it framed as a governance and financial matter, not just a product detail.
Isn’t this level of detail more relevant to larger, later stage companies?
The risk starts earlier than most founders assume, because even early stage products can carry heavy dependency on a single AI provider. I’d introduce the slide as soon as AI becomes material to the product, not once the company reaches a particular size.
How does VAQA’s board advisory work relate to AI exposure reporting?
It’s one of the specific practices I help boards adopt under my Board Advisory pillar at VAQA, alongside broader governance and reporting standards. I work with boards to build reporting habits that catch risk early rather than after it’s already caused a problem.
How often should the AI exposure slide be updated?
Every board meeting, in the same rhythm as cash runway and other standing financial slides. AI pricing and provider risk can shift quickly enough that anything less frequent risks missing a meaningful change.
What’s the biggest mistake companies make with AI reporting to their board?
Burying AI cost inside general opex without breaking it out, which hides both the trend and the underlying provider dependency from the people meant to be overseeing it. A board can’t manage a risk it can’t see.
Tom Batting is a Forbes 30 Under 30 entrepreneur, founder of Obby and Baluu, and founder of FirstMotion. He advises founders and leadership teams through VAQA.