Enterprise AI Doesn’t Have a Speed Problem. It Has a Measurement Problem.

By Mike Hollis, Co-Founder & President, Optura

Talk to enough enterprise leaders this year and you’ll hear the same board meeting stories. The same pressures. The same question.

Where are we on our AI, and what’s it returning? What’s the health of the investment, as a whole?

I’ve sat through this exact meeting more times than I can count. The CEO in the room isn’t frustrated because the team isn’t working. The team is working, the budget’s approved, the steering committee meets monthly, nobody can answer the question in plain terms. So the instinct is to lean harder on speed, shipping more use cases, standing up more pilots, moving faster.

That instinct is solving for the wrong variable.

Something shifted in enterprise AI conversations over the last few months. For two years, the boardroom question was velocity. How many use cases, how many agents, how fast can we scale this. Lately it’s become a conversion question instead. What value did all of this investment turn into?

Token spend and infrastructure costs are ballooning, with little productivity gain to show for it. Companies are pulling back spend as a result. That’s the headline version. Underneath it is a quieter, more useful realization. This was never a spending story. It’s a measurement story. Enterprises poured money into AI faster than they built any way to know what it was producing, and the bill just made that visible.

Boards aren’t wrong to ask the question. They’re finally asking the right one.

Here’s what leadership teams keep getting wrong: they assume the fix is more resourcing, or a stricter governance framework, or both.

Walk into almost any enterprise health plan or provider system today and both already exist. The budget’s there. So is the Center of Excellence, the steering committee, the CAIO, the review process. And the whole thing still moves at a crawl, because none of those structures were built to answer the one question the board cares about. Is this working, and where?

A fully funded, fully governed AI program can still tell you nothing about whether it’s healthy. Budget and governance tell you the program exists and is being watched, not whether it’s producing anything anyone can defend upstairs.e-day job that AI can do in two hours. Those fears are legitimate. Left unaddressed, they don’t go away. They go underground and quietly sabotage adoption. The honest answer is that the goal is upskilling people, not replacing them. That answer has to come from leadership, not a slide.

A program can look great on a dashboard, pilots launched, agents deployed, tokens consumed, while the organization underneath it is struggling. Here’s what determines whether that speed was worth anything.

Organizational readiness. This is the one I spend most of my time on with customers, and it’s rarely a data problem first. It’s a decision-rights problem. Who actually owns the call the AI is supposed to inform, and does that person have the standing to act on it? A team can have clean data and a working model and still stall here, because nobody redrew who’s accountable for what once the tool showed up. None of that shows up on a readiness checklist, and it decides whether an initiative gets used or quietly shelved.

Is governance built to protect the enterprise, or built so tight that nothing can ship?

The roadmap needs an actual sequence tied to strategy, not a pile of disconnected pilots competing for the same attention.

Use cases sized and bundled around value. Not one-off experiments nobody can defend in year two.

And returns. Someone has to be able to put a number on this that survives a CFO’s second question.

Miss any one of those, and speed just gets you to the wrong place faster.

It’s the argument Optura was built around. Enterprise AI doesn’t fail from a lack of ambition. It fails from treating implementation speed as the only thing worth tracking, while readiness, governance, roadmap, use-case discipline, and return all go unmeasured until a board meeting forces the question.

Looking at the whole system, not just how fast it moves, is the difference between a company that can defend its AI spend at the next board meeting and one still explaining, for the sixth time, why it isn’t moving fast enough.

Speed is easy to point to. Health is what determines whether any of it was worth it, and the programs I trust most are the ones that earned the room’s trust before anyone asked them to, not just the fastest ones.

See how Optura builds the roadmap and ROAI™ behind a healthy enterprise AI system.