# Journal

Canonical: https://www.meshqu.com/blog

Essays, papers and news on decision receipts, evidence and agentic decisions.

- [The payment is only one decision](https://www.meshqu.com/blog/agentic-payments-connected-decisions): What Mastercard RiskX taught me about agentic payments, payment fraud and why evidence, authority and judgement need to stay connected.
- [Opening the black box of autonomous work](https://www.meshqu.com/blog/opening-the-black-box-of-autonomous-work): We connected MeshQu to selected decisions across our coding task lifecycle. The records helped us identify bottlenecks, question our measurements and begin improving both the work and the rules around it.
- [When an AI agent acts, what should you be able to prove?](https://www.meshqu.com/blog/agentic-ai-governance): A general statement about your AI governance process cannot answer why one payment was made, and neither can a test showing the agent usually behaves as expected. You need the evidence behind that particular decision: which policy applied, what was considered, what the checks returned and who approved it.
- [Europe, the UK and Singapore want the same record. What would it have to be?](https://www.meshqu.com/blog/three-regulators-one-unanswered-question): An EU policy process, UK public opinion and a Singapore legal working group have reached for the same requirement in three different vocabularies: the ability to establish, after the fact, what an automated system decided and on what basis. They agree on the need. None has yet said what form the answer should take.
- [MeshQu selected to participate in the FCA Digital Sandbox](https://www.meshqu.com/blog/meshqu-selected-for-fca-digital-sandbox): MeshQu has been selected for the FCA Digital Sandbox. We will test a single claim against synthetic authorised push payment fraud data: that a challenged decision can be answered from a record made at the time, not reconstructed months later.
- [If two AI models read a policy the same way, will they decide the same way?](https://www.meshqu.com/blog/same-reasoning-different-decisions): We broke the previous experiment apart to find which piece was doing the work, then ran the same records through a second AI from a different company. The two models read the policy the same way and committed to verdicts on very different fractions of records.
- [What actually made an AI agent commit to a decision?](https://www.meshqu.com/blog/what-finally-changed-the-ais-mind): We ran the same AI over the same 283 UK procurement records five times, adding a layer of governance context each round. The decisive change was not more rules. It was being shown past decisions on similar cases, and that lever moved the agent in a single jump.
- [Why wouldn't the agent say no?](https://www.meshqu.com/blog/why-the-agent-wouldnt-say-no): An AI and a written rulebook reviewed the same UK government purchase records at the same moment. The AI named the same problems the rulebook did and almost never put a verdict on them. Changing the wording of one rule changed how often they agreed.
- [The AI Act keeps moving. What happens to the decisions already made?](https://www.meshqu.com/blog/ai-act-still-moving-decisions-stand-still): The EU AI Act has been delayed again and its timelines are still being argued over. The decisions already made under it cannot move with them.
- [When a model explains a decision, what has it proved?](https://www.meshqu.com/blog/explainability-is-not-proof): AI explainability shows how a model behaved. It does not show which policy, threshold and model version were in force when the decision was made.
- [Which policy was in force when the decision was made?](https://www.meshqu.com/blog/policy-as-code-still-doesnt-solve-this): Policy-as-code evaluates a decision against the rules in force now. Showing which rules applied six months ago means capturing the decision as it is made.
- [When the FCA asks why one payment was cleared, what can you show?](https://www.meshqu.com/blog/the-problem-isnt-policy): A UK retail bank does not lack AI policy. It lacks decision proof for one cleared payment when an FCA review picks it out of a sample eight months later.
- [What does a decision look like when you can prove it?](https://www.meshqu.com/blog/what-a-decision-looks-like): A core banking platform reconstructs decisions after the fact; a provable decision is captured at execution instead, as a single signed artefact.
- [Why can't your logs explain the decision six months later?](https://www.meshqu.com/blog/from-archaeology-to-proof): Logs usually record what happened, rarely why. Decision proof means capturing the decision itself at execution rather than reconstructing it from fragments later.