Anthropic's finance agents: what EU fintechs can adopt now

On May 5, 2026, Anthropic released ten ready-to-run agent templates for financial services — five for research and client coverage (pitch builder, meeting preparer, earnings reviewer, model builder, market researcher), five for finance and operations (valuation reviewer, GL reconciler, month-end closer, statement auditor, KYC screener). Each ships as a plugin for Claude Cowork or Claude Code, or as a Claude Managed Agent running autonomously on Anthropic's platform. Around them: Microsoft 365 add-ins, eight new data connectors, and Claude Opus 4.7 as the recommended model. The templates live on Anthropic's financial services marketplace.

The deployment choice is half your compliance footprint. Plugin mode behaves like desktop software with an LLM backend — the analyst stays in the loop and your existing audit trails apply. Managed Agent mode holds credentials, makes tool calls, and persists state on a third-party platform, which expands your DORA Article 28 vendor surface before you've shipped anything.

Our reading of the ten, by regulatory weight: pitch builder, meeting preparer, and market researcher are clean for EU adoption today — they produce drafts for humans, well below the AI Act's high-risk threshold. Earnings reviewer, model builder, and valuation reviewer are usable but need the human-review boundary designed explicitly — enforced thresholds, methodology versioning, quarterly re-validation. The last four deserve real pre-work: a stalled month-end close is a reportable DORA incident with a 4-hour notification clock, and a KYC screener that feeds creditworthiness decisions pulls the downstream system into AI Act Annex III high-risk territory. Watch the screener's escalation rate from day one — a falling rate is drift, not productivity.

When to move: adopt the first group now if coverage prep is a genuine bottleneck — six to eight weeks including connector and audit work is a fair scope. Wait if you don't yet have decision logs, data contracts, or rollback infrastructure; agents added to that stack accumulate compliance debt at feature speed.

Our scope is the engineering around the templates — connectors, audit instrumentation, EU-region inference, SIEM bridges — not the formal AI Act high-risk provider role. If your build needs one, we'll say so and point you to the right partner shape. The project planner is the easiest way to start, or see how we frame fintech engagements.