Founder, CEO, & Author | Building the Contributive Value (CV) platform — the human & agentic accounting layer for AI-era workforce. Currently in private beta.
The big four AI providers just committed over $9B to the exact same idea.
The new pitch? Stop selling software. Start sending armies of people to build it inside your company.
July 2: Microsoft launches Microsoft Frontier Company. $2.5B. 6,000 experts embedded in customer teams.
June 30: AWS puts $1B behind the same play.
May: OpenAI stands up a $4B deployment company via private equity.
Anthropic: Partners with Goldman Sachs & Blackstone on a $1.5B venture to embed engineers in mid-sized firms.
The model isn’t new—Palantir proved it two decades ago. What is new is that the entire market now agrees: the tool is no longer the value. The value lives in the brutal human work of making AI function inside real workflows with real data. They are spending billions to own that work. But a single word buried in this agreement decides everything: Which people?
The $9B is funding "deployer" humans. They arrive, build, prove a vendor-defined outcome, and leave. They are measured on throughput, cost, and adoption speed. They are very good at that. But that is not the same as being good for your workforce.
Nobody is spending a dollar on the second group: the incumbent workforce. The people who stay after the engineers pack up. When the deployment team leaves, you own a running system and a workforce whose contribution has quietly changed—and whose value nobody re-measured. The deployment wave doesn’t close this gap. It opens it. At scale.
Here is the question none of that $9B is built to answer: Once the AI is running, what is each person actually contributing, and how do you pay for it?
The reason providers won't answer is structural. A vendor whose revenue grows every time AI replaces a task cannot be the neutral party that tells you what your people are still worth. The incentive points one way: more adoption, more substitution. A measurement of human worth is only credible when the party doing the measuring doesn't profit from the answer.
There is a second structural problem keeping GCs awake: Access.
To make AI work, vendor engineers spend 6–12 months inside your data, workflows, and trade secrets. This same vendor serves your competitors. A confidentiality clause governs what they can do with what they see; it does not govern what they learn. Tacit knowledge walks out with every engineer and into the next engagement. Your proprietary edge becomes the raw material for their next product.
The open ground here isn’t deployment. Deployment is now the most crowded market in enterprise tech.
The open ground is control. Control over what your people are worth once AI arrives, and control over what your company gives away to get it there. Both are yours to hold. Both are easy to surrender by default.
You don't need to hire anyone to hold that ground. You need to ask better questions before you sign. Bring these six questions to any deployment team, regardless of the logo on their badge:
1. What is your definition of success? If the answer is adoption, usage, and cost, ask what any of it measures about the value of our people. (The answer will be nothing.)
2. What is the technical guarantee? We need a binding, verifiable guarantee that our data won't train your models or shape how you serve competitors. A confidentiality clause is a promise. Ask what happens to the tacit knowledge your engineers carry in their heads to their next client.
3. What is the measurement delta? When your team leaves, what will we be able to measure about our own workforce that we cannot measure today?
4. Who owns the aftermath? Six months after you are gone, how do we set pay and design roles for people whose work your system fundamentally changed? Who owns that problem?
5. Who owns the contribution data? Who owns the data about our employees' contribution that this engagement produces? Us, or you?
6. Where is the automation baseline? What incentive do you have to tell us when a role should be kept rather than automated?
If a vendor has clean answers to those six, sign with confidence. If they land as an afterthought, you’ve found the gap. It sits between the running system they leave behind and the company that has to keep operating, and competing, next to it.
The deployment teams are coming. The only real question is whether you will know what your people are worth, and what your company gave away, before they get there.
Published this morning on my Substack: Minnesota Has Every Data Center Advantage — and Is Losing the Race.
Cool climate. Abundant water. Strong grid. The 2040 Carbon-Free Electricity law that some hyperscalers are actively seeking out. The Minneapolis Fed identified the Upper Midwest as an emerging national corridor for AI data centers with Minnesota at the center.
So why did the AWS Becker project — multi-billion dollar investment, 15,000 construction job-years, hundreds of permanent jobs near a closing coal plant collapse? Why is Farmington in litigation? Why has Oppidan pulled back hyperscale plans in North Mankato and Hampton because Minnesota dropped in priority?
This isn't an environmental story. It isn't an economic story. It's a process story and the fix is community-first engagement infrastructure that doesn't yet exist.
The full piece is on Substack : https://t.co/K5PgTBNN1F
@jamil hello Jamil, we have a family connection. I sent you a detailed description. Please check your LinkedIn in Mail when you get a chance. Rod Brown