A fixed-scope read on what your AI actually costs, what it returns, and who owns it — delivered before you commit to anything larger.
Ask three people in most organizations what the AI spend returned and you get three different answers, none of them evidenced. The tools arrived through a dozen doors — a corporate card here, a bundled vendor feature there, a pilot that quietly became production. Finance sees the invoices. Engineering sees the systems. Nobody sees both.
That's not a governance failure. It's a measurement failure, and it's fixable in two weeks — which is roughly how long it takes to inventory what's running, score what it returns, and put a name next to each one. The hard part isn't the analysis. It's that no one inside the organization has both the mandate and the neutrality to do it.
Kickoff with you and IT, a document request, and four to six 30-minute interviews with department leads. Draft inventory back by day 5.
Workflow scoring against measured time and cost, shadow AI mapping, and vendor terms checked for training and retention language.
A 60-minute executive readout, keep/expand/kill decisions made in the room, and every artifact delivered in editable form.
Most AI engagements in this market start around $25,000 and run twelve weeks before you learn anything.
Find out what you actually need for $3,500 first. If the answer is a large build, you'll walk into that conversation knowing which three workflows justify it. If the answer is that you don't need one, that's the cheapest $3,500 you'll spend this year.
Thirty minutes. What's on your desk, who's asking, and whether this is even the right work. No deck.
Three candidate start dates and the fixed fee in writing before anything begins.
Interviews, evidence, scoring. About five hours of your team's time in total.
Sixty minutes with your leadership. Decisions made in the room. Every artifact delivered editable.
If the Diagnostic concludes that you don't need further help, it will say so in writing.
I'd rather lose the follow-on engagement than sell you one you don't need. An advisor who can't tell you "no" isn't giving you a read — they're giving you a quote.
The AI Value Diagnostic is $3,500, fixed fee and fixed scope, delivered in two weeks. Half reserves the window, half on delivery of the readout. There is no hourly billing and no change order.
Afterward there may be nothing. The Diagnostic is designed to stand alone — you get the inventory, the scorecard, the keep/expand/kill list and the 30/60/90 plan, and your team can execute all of it without me. If you want help with what comes next, we scope that separately.
I've run the infrastructure, not just studied it. Ten-plus years in hyperscale operations at AWS and Meta — deployment, power, capacity, and the AI/GPU buildout as it actually happened, including the parts that went badly.
Practically: a firm sends you a framework and a team of analysts. You get me, and answers grounded in having made these decisions with real consequences attached.
They get paid when the answer is "yes, build it." I don't.
An implementation firm is often the right call — there are good ones in this market and I'll say so. But you cannot get an independent read on whether your AI spend is working from the company that gets paid to sell you more of it. I'm the second opinion before a six-figure signature, and I have no product on the other side of the recommendation.
Yes. No vendor referral fees, no reseller relationships, no partner commissions. I have no financial interest in which platform you choose.
That's the entire point. If I made money on the recommendation, the recommendation would be worth nothing.
Maybe not. Here's the test: can your current policy tell you, today, how many AI systems are running in your organization and who is accountable for each one?
If yes, you're ahead of most. If the honest answer is "roughly," that gap is exactly what this work closes.
No, but it is for organizations with real scale, real spend, and real consequences. If AI is a line item somebody is now being asked to defend, the work applies.
If you're still experimenting and nobody's asking hard questions yet, save your money and call me when they start.
No. I'm based in Charlotte and I'll come to you if you're in the region — the interviews go better in a room. But the Diagnostic was built to run remotely and most engagements do.
Location changes the logistics. It doesn't change the work or the price.
No. The Diagnostic runs on interviews, spend exports, and vendor documentation you already have. No credentials, no software to install, no agents deployed in your environment.
Total ask on your side is roughly five hours across two weeks.
Anything proprietary to a former employer — internal metrics, timelines, vendor terms, or non-public strategy. Those obligations are permanent and I hold them without exception.
Industry dynamics and general practice are open ground, and that's where the value is anyway. Someone who'd break confidence for a prior employer would break it for you.
Tell me what's on your desk. If there's a fit I'll send three dates and a one-page scope. If there isn't, I'll tell you that instead.