Anthropic is not trying to convince Wall Street that AI is interesting anymore.
It is trying to package the work.
On May 5, Anthropic released ten ready-to-run agent templates for financial services. The list is exactly what you would expect if you have ever looked at how much finance work is repetitive, document-heavy, spreadsheet-heavy, and still weirdly manual:
pitchbooks, meeting prep, earnings reviews, model building, market research, valuation review, ledger reconciliation, month-end close, statement audit, and KYC screening.
Not glamorous.
Extremely monetizable.
The template matters
The word “template” is doing a lot of work here.
Financial firms do not want a cute chatbot that can summarize earnings calls if someone babysits it for twenty minutes.
They want repeatable workflows.
They want governed access to data. They want tool permissions. They want audit trails. They want something that can work inside Excel, PowerPoint, Word, Outlook, research platforms, market data systems, and internal repositories without turning every task into a new integration project.
Anthropic says each template combines skills, connectors, and subagents, with deployment as plugins in Claude Cowork or Claude Code, or as cookbooks for Claude Managed Agents.
That is the whole game.
The agent is not just the model. The agent is the packaged workflow around the model.
The office suite is the battlefield
Claude working across Microsoft 365 matters more than it looks.
Finance work moves through Excel, PowerPoint, Word, Outlook, PDFs, filings, transcripts, data rooms, research feeds, and internal notes. The work is not contained in one app. The context has to survive the jump from a model to a spreadsheet to a deck to an email.
Anthropic is explicitly aiming at that handoff.
Start a model in Excel. Carry the context to PowerPoint. Draft the deck. Prepare the note. Keep the human in the approval loop.
That is not science fiction.
That is what junior knowledge work already looks like, except slower and more fragmented.
Finance is the right first target
Finance is a natural place for this because the work is high value, text-heavy, number-heavy, compliance-heavy, and obsessed with process.
If an agent can save hours on pitch prep or month-end close without breaking controls, the ROI does not require a philosophical debate.
It is time saved. Faster review. Fewer manual passes. More analysis before the meeting. Less format-chasing after the meeting.
The jobs question will not go away. It should not.
But the first commercial wave is probably not “AI replaces finance.”
It is “AI eats the parts of finance everyone complains about and managers quietly measured anyway.”
That is how adoption usually starts.
The signal
The broader signal is that Anthropic is choosing vertical depth.
Claude Code proved that AI could become a real tool when the workflow was narrow enough and painful enough. Finance has a similar profile: expensive humans doing structured work in messy environments with too many documents and not enough time.
So Anthropic is not just selling intelligence.
It is selling a shape for the work.
That is the part to watch.
Because in enterprise AI, the winning product may not be the assistant that can talk about everything.
It may be the agent that knows exactly how a pitchbook, model review, KYC packet, or month-end close actually gets done.