AI Operations Management
The founding textbook for a new discipline. AI is a metered operating resource, not a software subscription, and it needs a discipline to run it.
The problem
Most organizations still buy AI the way they buy software: a seat, a license, a line in the IT budget. But AI doesn’t behave like software. Every question asked, every document read, every answer drafted is metered. Cost follows use, and use grows the moment something works.
That changes who has to manage it and how. The tools organizations already have, procurement for software and FinOps for cloud, each cover a piece. None of them treats AI as what it is: an operating resource that has to be sourced, metered, attributed, planned, routed, and justified.
What the book is
AI Operations Management is a textbook for that missing discipline. It sets out the core ideas in six parts of the job:
- Sourcing: where the capacity comes from, and on what terms
- Metering: measuring what is actually used
- Attribution: tying cost back to the work and the people who drove it
- Planning: budgeting for demand that grows with success
- Routing: sending each task to the model that does it well enough, for the least
- Value: proving the spend paid for itself
Fifteen chapters, in four parts. Every claim is either cited to a source or formalized as a definition or theorem. If it can be neither, it comes out.
Who it’s for
Leaders who have to make AI pay for itself, students in a course, and anyone who wants a clear account of how AI is actually run inside an organization.
Where it stands
Three chapters are published free on the web, each with a downloadable PDF. Print and ebook editions are in preparation. Its companion, the Applied AI Case Study Library, tests the book’s ideas against what has really happened.