Broadridge joins Anthropic’s Project Glasswing

Broadridge Financial Solutions entered an agreement with Anthropic to gain enterprise access to Anthropic’s generative AI models and governance tools for evaluation and potential integration into its client services.

The companies did not disclose financial terms or a public timeline. The arrangement allows Broadridge to test models and accompanying governance controls in environments that handle regulated and sensitive information.

Under Project Glasswing, Anthropic provides participating organizations controlled access to its models and tools designed to help manage safety, privacy and compliance risks. Controls include prompt and output monitoring, data handling restrictions and internal testing protocols.

Broadridge said it will assess use cases such as drafting and summarizing shareholder communications, automating routine client inquiries, extracting information from regulatory filings and streamlining operational documents. Broadridge provides services for banks, broker-dealers, asset managers and corporate issuers and operates platforms for proxy voting, regulatory communications and settlement operations.

Anthropic was founded by former AI researchers and has released iterations of its Claude family of models. The company has prioritized safety research, running external reviews and building mechanisms intended to limit harmful or inaccurate outputs.

The agreement gives Broadridge a structured way to evaluate model behavior and mitigation techniques in financial workflows and gives Anthropic additional input from a provider that serves regulated market clients.