Anthropic Launches 10 New AI Agents for Finance and Bank Work

Anthropic is stepping up its focus on AI agents in a big way. After seeing strong adoption of tools like Claude Code and Claude Cowork, the company has now rolled out 10 new AI agents aimed at helping banks, insurance firms, and other financial organizations automate routine work. The announcement was made at an event in New York, and it quickly sparked conversation online—especially around whether these tools could eventually replace entry-level jobs in finance.

There’s also some concern about how this might impact major SaaS companies like Infosys and TCS. Earlier this year, when Anthropic introduced Claude Cowork, the market reaction was immediate. Several SaaS stocks dropped sharply, with losses exceeding $285 million. For many observers, it felt like a wake-up call about how quickly AI could reshape the industry.

According to Anthropic, financial services have become its second-largest revenue source, just behind the tech sector. Interestingly, this move comes right after the company launched a new AI-focused services venture, reportedly backed by Wall Street.

All of these new agents run on Anthropic’s Claude platform and can be tailored to fit different business needs. The lineup includes tools like the following:

• Pitch Builder
• Meeting Preparer
• Earnings Reviewer
• Model Builder
• Market Researcher
• Valuation Reviewer
• GL Reconciler
• Month-End Closer
• Statement Auditor
• KYC Screener

One of the key advantages is that these agents can plug into existing software systems without forcing companies to completely change how they work. The idea is simple: reduce repetitive manual tasks and free up time for more strategic work.

In practical terms, these tools can handle things like building pitch decks, reviewing financial data, checking KYC documents, and managing month-end processes. Tasks that were once handled by junior analysts could now be completed much faster with AI support.

Anthropic explains that each agent is built around three main elements. First are “skills,” which define how the agent performs a task. Then there are “connectors,” which give it access to the necessary data. Finally, there are “subagents,” smaller AI models that handle specific parts of a workflow.

Companies can also customize these agents to match their own internal processes, whether that involves risk policies, approval chains, or modeling standards. The goal is to make adoption feel seamless rather than disruptive.

There are two main ways to use these tools. They can run as plugins inside Claude Cowork or Claude Code, working alongside analysts in familiar desktop environments. For example, a pitch-building agent could generate an Excel model, create a PowerPoint presentation, and draft an email in Outlook—all in one go.

Alternatively, businesses can use Claude Managed Agents on the Claude platform, where tasks can run more independently. These are better suited for longer workflows, like handling multiple deals or running overnight processes.

Even with automation, Anthropic says humans remain part of the loop. Users can review, edit, and approve outputs before anything is finalized or shared externally.

The company has also expanded integrations, including support for Microsoft 365 tools and new data connectors. A Moody’s MCP app has been added as well, giving financial professionals direct access to familiar datasets and analytics tools.

These AI agents are now available through Anthropic’s financial services marketplace. They can be accessed as plugins in Claude Cowork or Claude Code (on paid plans), or as Managed Agents on the Claude platform, which is currently in public beta.

This launch builds on Anthropic’s earlier push into finance. Last year, it introduced Claude for Financial Services, followed by specialized plugins for areas like equity research, private equity, and wealth management.