2 May 2026Briefing
Operating Models and What Success Looks Like
Why enterprise AI programs succeed or stall, and how to tell which is happening to yours.
Topic
Case Studies
Named or pseudonymous deployments analysed in depth, what was built, what worked, what didn't.
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
Operating Models & What Success Looks Like
Two enterprises with comparable AI ambition, similar vendor stacks, and similar talent pools routinely produce materially different results. The divergence almost never traces to the technology choice. It traces to the operating model, the design authority, build capacity, governance regime, run mo
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.
The Klarna walkback, what discipline gap looked like in numbers
In February 2024 Klarna announced an AI assistant doing the work of 700 agents. Eighteen months later it was rehiring humans. The numbers in between are a teaching case for the applied layer.