AI advisory for executive teams

Most companies can't answer six basic questions about their own AI.

I help executive teams find out where they actually stand — what's working, what's provable, and what the right next move is from here.

The six questions

Asked of each thing you've built or bought — separately, not as an average.

  1. What does one useful outcome actually cost here?
  2. Better than what — and how would you know?
  3. What happens when it's wrong?
  4. Would you notice if it quietly got worse?
  5. What does it know about your business that isn't written down anywhere?
  6. Who can stop it, and who can scale it?

The answers rarely match across deployments — and that mismatch is the finding. Most companies are advanced in one place and at zero in four others, which no single maturity score can tell them.

Who this is for

You've spent, and can't prove it worked

AI is deployed in several places. Spend keeps rising. Nobody can say what it returned, and the board has started asking. You need to know whether the problem is the technology, the people, or where you pointed it — because the fix is different in each case.

You haven't started, and don't want to waste the first year

You're being told to do something about AI. Every vendor has an answer. You'd rather skip the expensive mistakes the companies ahead of you already made — and know what to build first, and what not to buy yet.

How it works

01
A conversationForty minutes. I ask the six questions about your actual deployments. You'll know more about your own position by the end of it than an internal review would tell you in a month. No charge, no deck.
02
A diagnosticTwo weeks, fixed fee. The six questions run properly across your portfolio, ending in a document you can take to your board: what's provable, what to stop, what to fix first, and what it should cost.
03
Ongoing, if usefulA few hours a month as a standing advisor to the executive team. Two to four clients at a time, so the answer is sometimes no.

Who I am

I'm a practitioner, not a consultant. Eight years building AI and machine-learning systems — previously head of AI engineering at a major games publisher, currently CTO and cofounder of an agentic AI platform company in Austin.

That means I've paid the bills for a bad model-routing decision, set up the evaluation that told me it was bad, and had the argument about whether it was worth fixing. It also means I can tell when a vendor's claim is true, true-but-irrelevant, or quietly wrong — which is most of what this job turns out to be.

What I don't do. I don't implement, I don't resell, and an engagement never ends in a recommendation to buy anything — including from the company I run. If you need to build something after we talk, I'll point you at people who do that, and it won't be me. The advice is only worth paying for if nothing is riding on the answer.

Start with the conversation

If you can't answer three of the six, it's worth forty minutes.

sid@sidadvisory.com