Anthropic on May 5, 2026 unveiled ten production-ready Claude agent templates aimed at the most labor-intensive workflows on Wall Street, from drafting pitchbooks to closing the books at month-end. The announcement, made at an invite-only financial services briefing in New York, was paired with full Microsoft 365 integration and a Moody's-native app inside Claude that exposes credit ratings on more than 600 million companies.
The release positions Anthropic squarely against generic chatbot use cases and toward packaged, governable work — a deliberate shift from selling intelligence by the token to selling complete workflows that banks, hedge funds, and insurers can deploy in days.
Key Highlights
- Ten finance agent templates ship as plugins for Claude Cowork and Claude Code, and as cookbooks for Claude Managed Agents
- Microsoft 365 add-ins for Excel, PowerPoint, and Word are generally available; Outlook is in beta — Claude now keeps context across all four apps in a single session
- 17 data partners at launch: FactSet, S&P Capital IQ, MSCI, PitchBook, Morningstar, Chronograph, LSEG, Daloopa, Dun and Bradstreet, Fiscal AI, Financial Modeling Prep, Guidepoint, IBISWorld, SS&C Intralinks, Third Bridge, Verisk, and Moody's
- Claude Opus 4.7 leads the Vals AI Finance Agent benchmark at 64.37%, ahead of every competing model on the leaderboard
- Production customers include JPMorgan Chase, Goldman Sachs, Citi, AIG, Visa, Citadel, FIS, Carlyle, and Walleye Capital
The Ten Agents
Anthropic split the lineup into two functional pillars.
Research and Client Coverage:
- Pitch Builder — generates target lists, runs trading and transaction comparables, and drafts branded pitchbooks
- Meeting Preparer — assembles client and counterparty briefs from past notes, recent news, and talking points
- Earnings Reviewer — reads transcripts and SEC filings, updates models, and flags surprises versus the thesis
- Model Builder — constructs working financial models in spreadsheets from filings, data feeds, and analyst inputs
- Market Researcher — synthesizes sector developments, competitor moves, pricing, and broker research into memos
Finance and Operations:
- Valuation Reviewer — challenges every assumption in a valuation model against comparables and firm standards
- General Ledger Reconciler — matches books against bank statements and calculates NAV
- Month-End Closer — runs the monthly close checklist, prepares journal entries, and flags exceptions
- Statement Auditor — reviews financial statements for consistency, completeness, and audit-readiness
- KYC Screener — assembles entity files, reviews source documents, and packages escalations for compliance
Each template bundles three things: skills that hold the domain knowledge, connectors that govern access to data, and sub-agents that handle specific subtasks like comparable selection or methodology checks.
Microsoft 365 Becomes a Claude Surface
The Microsoft tie-in is the most consequential delivery channel. Analysts no longer need to leave Excel to pull comps, draft a slide, or summarize an email thread — Claude operates as a single agent across the suite, maintaining shared context between the workbook, the deck, the document, and the inbox.
Citadel's head of core engineering said Claude for Excel "meets analysts where they live in data and models, building coverage models with a step-change in efficiency." JPMorgan Chase CEO Jamie Dimon was quoted describing a Claude Code session that produced a complete financial dashboard in twenty minutes "with all the backup, and all the research, and it was very accurate."
Why It Matters
The launch confirms two trends Anthropic has been signaling for months. First, the company is moving up the stack — from API access, to managed agent infrastructure, to packaged vertical workflows that compete with traditional software vendors. Second, Wall Street is no longer evaluating AI; it is buying it.
Anthropic CEO Dario Amodei told the briefing audience the company achieved 80x annualized revenue growth in a single quarter — eight times the internal projection — characterizing the moment as "absolute radical uncertainty." Walleye Capital's CEO disclosed that 100% of its 400-person hedge fund now uses Claude Code daily. AIG reported that Claude reached 88% accuracy on insurance claims work, measured against human experts.
Background
The finance agents follow Anthropic's April release of Claude Managed Agents, the hosted runtime that lets companies build, sandbox, and observe long-running agents without operating their own infrastructure. The new templates are the first verticalized application of that platform — proof that the same plumbing can be productized for a specific industry rather than left as a developer primitive.
FIS CEO Stephanie Ferris said Anthropic was "the clear choice" for the company's anti-money-laundering work, where a Claude agent now compresses AML investigations from days to minutes.
What's Next
Anthropic is signaling that more verticals will follow the same pattern — package the workflow, govern the data, sell the outcome. Legal, healthcare, and supply-chain templates are the obvious adjacencies. The 17-partner data marketplace, exposed through Anthropic's MCP standard, also creates a flywheel: every new connector makes the existing agents more useful, and every new agent makes the connectors more valuable.
For finance teams in the MENA region, the practical takeaway is that the entire research-to-close cycle now has off-the-shelf AI scaffolding, governed at the connector level and auditable end-to-end. The question is no longer whether to deploy AI agents on financial work — it is which agent to map to which job, and how fast.