You've watched the demos. You've seen the keynote. Now you have to use it on Monday. Anthropic's May 2026 launch put Claude inside Excel, PowerPoint, Word, and Outlook. Ten pre-built agents shipped on a marketplace. Moody's went native. Microsoft 365 integration went generally available. Every CIO in financial services has the deck on their desk and every analyst has a Slack message from a senior asking whether this changes their week. It does. But probably not in the way the keynote suggested. Claude for Financial Services is the practical guide the launch didn't give you. Fifteen chapters in four parts, written for the analyst, associate, or director who spends their day in Excel, PowerPoint, Word, and Outlook — and who needs to know how to get reliable work out of these tools, not the demo version. Part One sets up the stack: what changed in the launch and what didn't, and the five surfaces (Cowork, Claude Code, the Microsoft 365 add-ins, the marketplace) you need to know. Part Two covers the work. Excel models that hold up under questioning. PowerPoint decks that survive the page-turn. Word memos that read like your firm wrote them. Outlook triage and meeting prep. Research that's worth reading. Real prompts, real failure modes, real slash commands — used in the workflow they belong to. Part Three covers the agents. An honest review of the ten pre-built agents that shipped at launch — Pitch, Credit Memo, KYC Screener, Underwriting, Month-End Close, GL Reconciler, Statement Audit, Market Researcher, Earnings Analysis, Insurance Claims. The deployment-mode decision: plugin or managed, in your loop or on Anthropic's infrastructure. The configuration work that turns a generic agent into a firm-specific one. And the compliance question — with the FIS / BMO AML deployment as the anchor case study. Part Four is about judgment. The failure modes you need to know before something reaches a client. The staircase of autonomy and the collaboration patterns that produce good work. What this means for an analyst-to-MD career — the honest version, not the LinkedIn version. And a closing chapter on what to build next, this week, with what you have. Who this book is for. Investment banking analysts and associates building models and decks. Asset managers and equity researchers digging through filings. Credit, risk, and compliance teams running memos and KYC. Operations and finance leads closing the books. If you spend most of your day inside Microsoft 365 and Cowork, this book was written for you. What makes this book different. It is independent — not commissioned, endorsed, or affiliated with Anthropic. It is practical — every chapter teaches a method, walks a workflow, names the failure modes, and closes with concrete things to try this week. It is built to last — the tools will keep changing, but the discipline survives. And it is compliance-aware — the regulated workflows (KYC, AML, credit decisioning, statement audit) each get specific treatment, with the regulatory regimes named. About the author. Fransesco Malila writes practical and thought-leadership books on AI in professional contexts. He has had no commercial relationship with Anthropic in the writing of this book. The tools will keep changing. The method holds.
