The lab is not the business
Work still has to land in a record.
A model can recommend. Only a workflow can act. When the chat closes, the company still has to store that decision in an official record, name an owner, and see it in the P&L or with the customer. Until then the budget bought a lab.
Library
Recent entries
Insights answer a live executive question. Frameworks hold the language those articles can cite.
The Consulting Pyramid Has Collapsed. Can AI Services Become a Diamond?
AI hits the bottom of the consulting labor pyramid first. The question is whether a services firm can change shape, or only pass efficiency back to clients as price pressure.
Insight · Jun 22, 2026IN-04What Should PE Ask Before Buying an AI Services Firm?
AI readiness in services firms cannot be measured by demos or partner badges. Diligence should test revenue quality, delivery margin, and whether AI changes economics.
Insight · Jun 15, 2026FW-01Enterprise Software Layer Model
A practical model for placing enterprise AI work across systems of record, data platforms, semantic meaning, inference, orchestration, and governance.
Framework · Jun 14, 2026FW-03Inference vs Orchestration
Inference is the model producing an output. Orchestration is the logic around that call: what happens before, during, and after, including data access, tools, permissions, retries, and action.
Framework · Jun 14, 2026FW-04AI Governance and Control Layers
Governance should rise with operational consequence, not model novelty. Production AI needs control, evidence, and an owner inside the workflow.
Framework · Jun 14, 2026IN-02Systems of Record Are Losing the User Interface War
The system of record may remain authoritative, but AI agents, workflow platforms, and orchestration tools increasingly own where users work.
Insight · Jun 8, 2026Readers
One voice. Three rooms.
Enterprise leaders hear operating judgment: owners, records, write-back, and whether the customer feels the difference. Investors hear diligence questions that survive a management presentation. Builders and collaborators see a practitioner who has sold, implemented, partnered, and had to live with the result.
Frameworks
The models underneath the essays
Start with where authority lives, where judgment becomes action, and how much control the consequence requires.
FW-01
Enterprise Software Layer Model
A practical model for deciding where AI work belongs, which system stays authoritative, what commits action, and what evidence proves the result.
FW-02
Systems of Record After AI
The official record stays load-bearing. The action surface moves. Strategy has to fund the seam, not pretend the ledger will become fast.
FW-03
Inference vs Orchestration
Separate the model-serving question from the system question. Governance has to watch both, or it will log tokens and miss the business process.
Advisory
The public point of view has a practice behind it.
A seasoned practitioner does the work: strategy, diligence, partner field alignment, and implementation readiness. It is not a body shop.
Without customer impact and ROI, the technology discussion is mostly theater.
Why Diagenic?
In geology, diagenesis is the process by which dispersed material consolidates under pressure into coherent rock. The practice uses the same idea: fragmented expertise, partner motions, and delivery models have to become something load-bearing, or they stay theater.