Series
The AI Engineer Playbook
A playbook for engineers building enterprise AI and agentic systems: from model serving to an authorised business outcome, and the metrics that tell them apart.
The AI Engineer Playbook is for the people who build and run enterprise AI and agentic systems. Where the Agentic AI Playbook gives leaders the five rungs and the fifth-rung commit boundary, this series takes the engineer’s route through them: how a model is served and what the serving trade-offs are, how context is constructed and versioned, how answer quality, tool choice and action correctness are evaluated separately, how proposals become typed requests under delegated authority, what is checked before a change commits, how the whole workflow is traced, and how time, reliability and cost are measured per completed, authorised outcome.
The series draws on the Contextual Agentic Enterprise (CAE), Contextual Agentic’s architecture and operating model, and on engine documentation and original papers, which each piece cites with dates. Figures from papers are the authors’ own; the worked examples are fictional. Any adoption of CAE should be tailored to the organisation in consultation with Contextual Agentic or another advisory service. The complete patterns, the Context Contract and the evidence bundle are in the book.
In this series
-
The AI Engineer Playbook: From Model Serving to Authorised Outcomes
Eight engineering decisions take an agent's request from model serving to a completed, authorised business outcome: define the action and its budget, serve the model, construct versioned context, evaluate three things separately, propose typed actions under delegated authority, check before commit, trace the whole workflow, and measure per completed authorised outcome. Faster tokens shorten only the first clock.