2 May 2026Briefing
Evaluation and Governance
How to know enterprise AI works, and how to ship it safely. Operational practice, not slogans.
Topic
Evaluation & Observability
Eval frameworks, online evaluation, and observability stacks for LLM-backed systems running in real environments.
2 May 2026Briefing
Architecture and Retrieval
The patterns that distinguish production AI from demos.
2 May 2026Research
Trust in Enterprise AI: Evaluation as Practice, Governance as Delivery
Two disciplines determine whether enterprise AI earns operational trust: evaluation, the practice of measuring whether a system actually works in production; and governance, the delivery of policy as code, controls, and accountable workflows. Both remain underspecified. Evaluation in many organizat
2 May 2026Research
Production AI Architecture: Patterns That Distinguish Production from Demo
By 2026, enterprise AI systems are no longer differentiated primarily by which large language model they use. The frontier models, Anthropic's Claude Opus 4.7, OpenAI's GPT-5.2, Google's Gemini 3 Pro, are converging on capability for the median enterprise workload. What separates production-grade s
2 May 2026Research
The First Year of the Applied Layer: A Synthesis
The most consequential layer of the AI buildout is not the foundation models themselves but what sits between them and the organizations that deploy them: architecture, integration, evaluation, and governance. The public record has clarified the picture rather than settled it. The applied layer is
Morgan Stanley's evaluation framework, from 7,000 questions to 100,000 documents
Morgan Stanley shipped two assistants in eighteen months. The visible artefact in both cases was the model. The invisible artefact, the part that decided whether the rollouts compounded, was the evaluation harness underneath.