The gap in the AI consulting market
Search for AI consulting services and you find two camps. Big firms selling AI readiness strategy at enterprise prices, where the deliverable is a roadmap. And dev shops selling engineering hours, where the deliverable is code someone still has to turn into a business tool.
What almost nobody sells is the thing operating companies actually need: someone who understands how a business runs, its cadence, pricing, inventory, orders, reporting, and cash, and can build AI-powered systems around those workflows that people use every day.
The gap is not AI expertise or engineering talent. It is operators who can build.
The economics, with real numbers
Here is what makes this category worth your attention, from a system I built and still operate:
| Item | The old equation | The AI-native equation |
|---|---|---|
| Capability | 10+ separate SaaS subscriptions plus manual glue work | One custom operating system: 52 modules across accounting workflows, order-to-delivery operations, customer pipeline, and strategic reporting |
| Annual running cost | Six figures per year in licenses at enterprise SaaS pricing | Roughly $300 per month in AI and hosting costs |
| Fit to the business | 60% fit, workarounds for the rest | Built around the actual workflows, including capabilities no SaaS vendor sells at any price |
| Build cost | $280K to $650K to replicate with traditional development | A fixed-scope sprint, weeks not quarters |
Figures from a live system in production as of August 2026. Every business is different; part of the assessment is telling you honestly what your version of this math looks like.
A second, smaller example: a spec home builder replaced a $10,000+ per year project management platform with a construction-native tool built in days, one that matches how his crews actually work instead of forcing his crews to match the software.
What an engagement looks like
- Interview. I sit with you and your team and map how the business actually runs: the workflows, the handoffs, the spreadsheets nobody admits to, the reports someone builds by hand every Monday.
- Identify the leverage. Not everything should be rebuilt. We rank the workflows where AI-native systems change the economics, and skip the ones where your current tools are fine.
- Build fast. Working software in days and weeks, shipped in usable increments, tested against the real workflow with the people who will use it.
- Deliver and run. Your team runs on the system. You own it. Running costs stay in the hundreds per month, and the system evolves as the business does.
Pricing, published
| Engagement | Range | Typical |
|---|---|---|
| Build Sprint (fixed scope) | $15,000 to $40,000 | $25,000 |
| Embedded Operator (build and run) | $8,000 to $15,000 / month | $12,000 |
| Sounding Board (advisory) | $3,000 to $7,000 / month | $5,000 |
Written scope, weekly milestones, a 30-day out, and a target of 3x return on the fee within 12 months against a baseline we set before starting. More context on how these rates compare to alternatives is on the fractional COO cost page.
What this deliberately does not include
- No model training or ML research. I build on frontier AI platforms; I do not compete with them.
- No consumer app development. These are operating systems for running businesses, not products for app stores.
- No enterprise system-of-record migrations. If you need a large integrator to move your ERP, hire one.
- No AI readiness decks. If a workflow is worth doing, we build it. If it is not, I say so and you keep your money.
Being specific about what this is not is the fastest way to be trusted about what it is.
Who this is for
Founder-led businesses roughly $5M to $50M+, PE-backed portfolio companies, and search fund operators. Big enough that operational drag costs real money. Lean enough that the $400K enterprise software project was never going to happen anyway. If that is you, the math in Section 02 is the most interesting thing you will read this quarter.