Blockworks Lets Exchanges Take Compliance Alerts Directly From AI Agents

This reframes compliance monitoring as an agent-first workflow, with human review as a second layer rather than the default. The labelling of agent versus analyst output is the more consequential design choice: it gives desks a way to triage machine-generated alerts without pretending they carry the same weight as reviewed ones. Exchanges adopting this take on a new dependency — the speed of the alert now outruns the certainty behind it. Whether compliance teams treat unreviewed agent flags as actionable, or wait for the analyst pass, is the open question.