Sixty-one percent of firms in the financial services industry are already using or assessing generative AI, signaling a rapid embrace of the technology within a heavily regulated sector. While many anticipate content creation as the primary application, Anthropic reports that 56% of firms utilizing or evaluating agentic AI prioritize “knowledge management and retrieval” as the key use case. This focus is about efficiently accessing and utilizing existing information to drive better decision-making, rather than flashy outputs. Anthropic customers are demonstrating this commitment by deploying multiple Claude products, including Claude for chat and research, Claude Cowork, Claude Code, and integrations for Microsoft 365, to build comprehensive AI suites integrated into daily workflows. As executives face pressures on margins, talent, and regulation, they are embracing AI technology for faster, better answers.
Generative AI Adoption in Financial Services
Eighty-nine percent of financial services firms report that AI has increased revenue while lowering costs, a figure demonstrating rapid integration of the technology within a traditionally cautious sector. Beyond initial experimentation, adoption is deepening, with organizations increasingly focused on practical applications that leverage AI’s capacity for knowledge management and efficient information retrieval. The industry’s historical eagerness to embrace technologies offering a competitive edge in research and productivity is visible in the current wave of AI implementation, though the emphasis is shifting from simple question answering to systems capable of independent action. This focus isn’t merely about automating routine tasks; it’s about unlocking the value embedded within vast internal datasets and external information sources.
At Anthropic, this trend is readily apparent. “Claude has changed how teams get answers and Claude Code has transformed how engineers build software,” reflecting a broad impact across different functional areas. Organizations aren’t limiting themselves to a single AI product; the majority are deploying multiple Claude applications in tandem. This multi-product approach allows firms to build a comprehensive AI suite integrated into existing workflows. The Claude Platform, Anthropic’s API, is enabling quantitative teams and platform engineers to embed AI directly into core systems like trading platforms and risk assessments. Anthropic’s Claude Managed Agents offer a cloud-hosted solution for deploying agents programmatically, exemplified by a “Valuation Reviewer agent handling month-end reconciliation across a multi-hour close.” Customization is further facilitated through connectors, skills, and plugins, allowing firms to tailor AI capabilities to their specific data requirements and regulatory constraints. These building blocks enable access to critical data sources, encode repeatable task workflows, and bundle functionalities into installable packages, ultimately driving efficiency and consistency across complex financial operations.
Claude Cowork for Multi-Step Financial Operations
The financial industry’s embrace of artificial intelligence has moved beyond simple experimentation, with a substantial 61% of firms now actively using or assessing generative AI technologies to reshape operations and improve performance. This emphasis reflects a pragmatic approach, prioritizing access to existing information over the creation of novel outputs, and is particularly evident in how companies like Anthropic are seeing their products deployed. Anthropic’s suite of Claude products is increasingly being adopted as a cohesive unit, rather than isolated tools. This multi-product strategy allows for a deeper level of integration into existing workflows, extending beyond basic tasks to encompass complex, multi-step operations. Claude Cowork, specifically, is designed to address these intricate processes, functioning as a desktop application capable of interacting with local files and connected applications.
It can, for example, read and write local documents, call connected services such as Google Drive, Gmail, Box, and Asana, and coordinate multi-step work across files and apps. The application of Claude Cowork within financial services is diverse, spanning research, deal work, analysis, claims processing, model reviews, and reporting. A concrete example of its utility is the ability to build a pitch book, reconcile a month-end close, or prepare a board packet. This isn’t simply automating routine tasks; it’s about augmenting human capabilities and streamlining complex projects. Anthropic offers the Claude Platform and Claude Managed Agents for firms seeking to build and deploy custom applications, allowing for even greater tailoring to specific needs. The company has also pre-built reference architectures for financial services, providing templates for tasks like pitch book creation and general ledger reconciliation.
As AI becomes the interface for decisions, trust becomes the standard. Moody’s decision-grade connected intelligence is key to unlocking AI for high-stakes credit and compliance decision-making.
Rob Fauber, President & Chief Executive Officer, Moody’s
Claude Code for Quantitative Finance & Engineering
Quantitative researchers are leveraging Anthropic’s Claude Code to accelerate the migration of legacy COBOL and Java systems, a task traditionally requiring significant time and specialized expertise. This move exemplifies a trend within financial institutions: 61% of financial services firms are using or assessing generative AI, with 89% reporting that AI has increased revenue while lowering costs. While much attention focuses on AI’s potential for generating reports or drafting communications, the core value proposition for agentic AI in finance lies in efficient knowledge management and retrieval, cited as the top use case by 56% of firms currently experimenting with the technology. Claude Code distinguishes itself as a command-line interface specifically tailored for technical teams, offering capabilities that extend beyond conversational AI. Finance firms are deploying it to write and audit quantitative code, including crucial trading logic and data pipelines, all while maintaining strict version control.
The platform’s ability to operate directly within a terminal or integrated development environment (IDE) allows for seamless integration into existing workflows, minimizing disruption and maximizing efficiency. Anthropic’s strategy isn’t centered on a single product, but rather a suite of integrated AI applications. Financial services firms are utilizing these agents programmatically, for example, deploying a Valuation Reviewer agent to automate month-end reconciliation processes.
We achieved full ROI within the first three months, through productivity improvements, speed-to-market gains, and agency cost reductions. In some use cases, productivity increased by 100%, beyond my wildest expectations.
Olga Pirog, Global Head of Data and AI Transformation, IG Group
Microsoft 365 Integration Enhances Analyst Workflows
The demand for efficient data analysis within financial institutions is driving rapid integration of artificial intelligence directly into existing software ecosystems. Beyond simply adopting AI tools, firms are increasingly focused on embedding these capabilities into platforms like Microsoft 365, streamlining workflows for analysts and operators. This isn’t about replacing human expertise, but rather augmenting it with systems capable of rapidly processing and synthesizing information. A NVIDIA survey found that 61% of financial services firms are using or assessing generative AI, with 89% reporting that AI has increased revenue while lowering costs. A key driver of this integration is the need for improved knowledge management and retrieval. Anthropic reflects this trend, observing that customers are deploying multiple Claude products in concert.
This multi-product strategy extends to Claude Cowork, designed for project-level work across files and applications, and crucially, Claude for Microsoft 365, which addresses the significant time analysts spend within Excel, PowerPoint, and Word. The power of the Microsoft 365 integration lies in its contextual awareness. The platform allows Claude to search and analyze content within Outlook, Teams, SharePoint, and OneDrive, not just from chat interfaces or uploaded documents. This means a financial model built in Excel can be seamlessly transformed into a presentation deck or client email without requiring manual rebuilding of the information. An example of this capability is the ability to “update this DCF with the new guidance and turn it into a three-slide deck.” This focus on interconnected AI tools, rather than isolated solutions, is becoming the norm.
These architectures package skills, tools, and data connections into readily deployable templates, accelerating the implementation process and ensuring consistency. Ultimately, the goal is to create a cohesive AI ecosystem that empowers analysts to focus on higher-level strategic thinking, while automating routine tasks and unlocking deeper insights from complex data sets.
Claude Platform & Managed Agents for Custom Applications
This approach moves beyond standalone tools; organizations aren’t simply adopting AI, but constructing a suite of applications. The ability to customize Claude through connectors, skills, and plugins is central to this strategy. Skills, described as “reusable, encoded workflows,” provide consistency crucial in regulated industries like finance. These skills can range from simple Markdown instructions to complex scripts, allowing firms to standardize tasks like credit memo creation or KYC onboarding checks. The Claude Managed Agents service further streamlines this process. Rather than managing the underlying infrastructure, firms can deploy custom agents as a hosted cloud service, with Anthropic handling runtime, permissions, and audit trails. Anthropic explains that “a firm builds an agent on the platform and Anthropic runs it as a cloud service, with the production scaffolding Managed Agents provides.” This is exemplified by the potential to deploy a “Valuation Reviewer agent” programmatically, automating month-end reconciliation with detailed execution tracing.
The combination of the platform and managed agents allows firms to move beyond experimentation and into production deployments, embedding AI directly into core business processes. The company highlights that firms can use the platform to “add Claude to our advisor platform to draft personalized portfolio commentary at scale, with a human in the loop.” This granular level of integration represents a significant step beyond simple chatbot implementations, promising to reshape how financial institutions operate and deliver value.
