Start a conversation Contact

Research & insights

The thinking, published before it is billable.

What we have worked out on engagements and had no reason to keep. Written for practitioners: the method is visible, the uncertainty is stated, and where something we published has aged badly we say so rather than quietly deleting it.

Latest · Governance · 6 min

What the EU AI Act actually asks of a retrieval system

Most teams building retrieval-augmented systems are preparing for the wrong obligations. The Act is not primarily interested in your model — it is interested in whether you can reconstruct, months later, why a particular answer was given.

19 Aug 2026 Read →

25 Aug 2026 The benchmark you are procuring against is a dataset nobody audited Experts re-checked two widely used text-to-SQL benchmarks and found annotation errors in more than half the examples of each. Correcting the labels changed the ranking of the agents measured against them, which is the part that matters to anyone selecting a supplier on a leaderboard position. Data 23 Aug 2026 There will be no draft of the ICO's agentic AI guidance The ICO has agentic AI guidance in drafting for Winter 2026 and no public consultation on it, so no draft will circulate for anyone to argue with. What it will land on is already in print — the regulator has said that design and architecture determine how data protection law applies to an agentic system. Governance 22 Aug 2026 The first AI Act standard is published. Presumption of conformity is not. EN 18286 reached publication in July 2026, and publishing a European standard is not the same act as citing it in the Official Journal. What it settles is which records a high-risk provider will be asked for — and a quality management system is the one obligation that cannot be assembled after the fact. Governance 21 Aug 2026 The AI Act Omnibus deferred the classification, not the architecture The Digital Omnibus on AI moved the high-risk obligations to December 2027 and August 2028 and left Article 50 running from 2 August 2026. It attaches to how a system is built and surfaced rather than to what it is used for, which is why an ordinary enterprise assistant sits inside the deadline that did not move. Architecture 12 Aug 2026 Establish the statistical baseline before you buy a GPU A dull regression, run first, is the cheapest insurance policy in applied machine learning. It either tells you the expensive model is unnecessary, or it gives you the only number that can prove the expensive model was worth buying. Method 28 Jul 2026 Five decisions, not two: how an assessment should end An assessment whose only possible conclusions are "approved" and "not approved" is not an assessment. It is an approval process with a report attached, and everyone in the room knows which answer is expected. Governance