Blog

  1. A father's 80,000 notes, and the records your company keepsThe model is rented and getting cheaper. What a company should own instead: its records, the tests for good work, and models trained on both.
  2. Where AI fits in a retailer's queuesWhere an AI system fits in a retailer: returns, reorders, order questions and markdowns, prepared from your data within limits your team sets.
  3. What Klarna measured, and what it missedKlarna's AI assistant handled 2.3 million chats in its first month. By May 2025 Klarna was recruiting people again. What to measure before launch.
  4. Most firms use AI. Few can show profit.McKinsey found 88% of organisations use AI, yet about 6% attribute over 5% of EBIT to it. The gap usually starts with a missing baseline.
  5. What AI can prepare for your lawyersWhere an AI system fits in a law firm: new-matter intake, conflict searches, bundle summaries and precedent drafts, each checked by a lawyer.
  6. Why your AI should run in your cloudWhat running an AI system in your own cloud account means for client promises, UK GDPR, data retention and changing supplier, with five questions to ask.
  7. What an AI system does with a claimHow an AI system prepares each new claim for a handler: policy found, cover dates checked, missing documents chased and every step recorded.
  8. How to know an AI system worksWhat evaluations and guardrails are, why you need both before an AI system handles customers or money, and four questions to ask whoever builds it.