How Claude is transforming financial services

The financial services industry is undergoing a fundamental shift from passive curiosity about AI to active production deployment. Firms are moving beyond simple chatbots to build and integrate sophisticated AI solutions into their core workflows. This transition allows analysts to spend less time on mundane, manual tasks and more time on high-value activities like building client relationships and deep business analysis.

For instance, the Norwegian Sovereign Wealth Fund (NBIM), which manages around 9,000 portfolio companies, has built its own integrations using tools like the Model Context Protocol (MCP). Their portfolio managers now query these integrations daily to gain insights into their portfolio companies, demonstrating a move toward deeply embedded AI.

From Code to Complex Financial Analysis

Claude’s proficiency in finance stems from its foundational strength in coding. While software engineers represent a small fraction of the global population, the logical reasoning, precision, and systems-thinking required for coding are highly transferable to other complex domains, especially finance.

Financial analysis involves interacting with complex, regulated digital systems where accuracy and auditability are paramount. Claude’s ability to understand and generate structured logic for code translates directly to its capacity for:

• Interacting with complex systems: Finance relies on a multitude of data platforms and software.
• Exposing its reasoning: Claude can show its work, which is crucial for verification and trust in a regulated field.
• Creating structured outputs: The same capability that allows Claude to write code enables it to generate pixel-perfect Excel models and PowerPoint presentations. The file creation feature, for example, operates by running Python code in a virtual machine to build and edit documents.

Anthropic’s founding principles in AI safety are also a cornerstone of its financial services offering. This focus on safety manifests in three key ways:

1. Secure Deployment: Ensuring solutions can be safely integrated into enterprise environments.
2. Accuracy and Fidelity: Training models to answer questions with a deep understanding of the problem domain.
3. Trust and Auditability: Providing users with the tools to verify and understand the model’s outputs.

Claude for Financial Services: An Agentic Framework

Claude for Financial Services is an end-to-end agentic solution built on a “retrieve, analyze, and create” framework, designed to cover the entire lifecycle of an analyst’s workflow.

• Retrieve: Large language models excel at parsing vast amounts of data. The goal is to connect Claude to all the core data sources analysts use—such as S&P, FactSet, and PitchBook—to uncover insights faster than competitors.
• Analyze: Beyond retrieval, the solution enables analysis at scale. This involves not just running calculations but also understanding core financial concepts and allowing analysts to inject their own judgment into financial models.
• Create: The final step is to generate client-ready and boardroom-ready deliverables. The system is designed to produce polished outputs in formats like Excel, PowerPoint, and Word, making it a complete, autonomous system.

This solution is built on three layers:

1. The Models: Anthropic is a research lab at its core, focused on making Claude the best model for finance. This involves close collaboration with early customers like BCI and Perella Weinberg to understand key use cases and identify gaps.
2. Agentic Capabilities: These are the product features that enable users to interact with the models, such as deep research tools and integrations that embed Claude into surfaces where analysts work every day (e.g., Excel, Chrome browser extension).
3. The Platform: A flexible platform with pre-built integrations with industry partners like S&P and FactSet, allowing for easy and powerful deployment.

Transforming Workflows: The BCI Use Case

British Columbia Investment Management Corporation (BCI) has transformed its competitive analysis (“comps”) workflow using Claude. Previously, analysts would manually create and update static Excel sheets comparing financial and operational metrics for various companies.

Now, BCI uses Claude’s Artifact feature to connect directly to S&P and FactSet data feeds. The result is a live dashboard where metrics can be compared in real-time. This dashboard can be updated with a simple prompt and is shared directly with managing directors. This represents not just an acceleration of work, but a fundamental transformation of the workflow itself from a static, manual process to a dynamic, interactive one.

Future Directions: Memory, Specialization, and Collaboration

The roadmap for Claude for Financial Services is focused on deepening its capabilities and integration within the industry.

• Memory and Personalization: A key area of development is building memory systems that allow Claude to maintain context and learn user preferences across different tools and sessions. For example, Claude could remember a user’s preferred formula for an EBITDA calculation or their preferred data source for specific metrics, acting like a “good intern” that improves over time.
• Deeper Specialization: Anthropic is investing in finance-specific pre-training and post-training for its models. The product team is also focused on tailoring workflows for specific sub-verticals, recognizing that private equity firms have different needs than hedge funds or investment banks.
• Ubiquitous Integration: The goal is to have Claude available everywhere analysts work, with a strong focus on improving the quality of outputs in Excel and PowerPoint.
• Partnerships and Customer Collaboration: Deepening partnerships with industry leaders like S&P and FactSet is crucial. Most importantly, the development process relies on close collaboration with enterprise customers. By translating customer needs into concrete evaluation sets (“evals”), Anthropic can directly incorporate these requirements into the model training and product development pipeline, ensuring the final product meets the real-world needs of financial professionals.

• Tags:
• AI Agents
• Large Language Models
• Feature
• Anthropic
• Claude
• Tool Use
• Financial Services