FactSet and Google Cloud have announced a partnership to bring Gemini models, Google enterprise search and Google Cloud infrastructure into financial-research and workflow tools used by investment professionals.
FactSet is a Connecticut-based financial data firm. The partnership is focused on three parts. FactSet plans to embed Google’s enterprise search and Gemini model capabilities in FactSet Workstation, connect FactSet’s financial intelligence to Gemini Enterprise through Model Context Protocol and agent-sharing functions and jointly develop agents for portfolio operations, deal advisory and corporate finance.
The companies said that the partnership aims to provide financial AI that is sourced and auditable. “Financial institutions require AI tools that anchor advanced technology in reliable, industry-specific intelligence,” Karthik Narain, chief product and business officer at Google Cloud, said in the announcement.
Missing details on audit evidence
The auditability claim shifts attention from the partnership announcement to implementation detail. FactSet’s release says the partnership targets “fully sourced, auditable, and defensible” outputs in regulated environments, but it does not disclose launch dates, named production users, pricing, performance results or the exact evidence users will receive when an agent produces an answer or takes a workflow step.
Moving beyond research to autonomous actions
Google’s release describes workflow agents as systems that can execute a sequence of steps or actions, including AI automation and human intervention.
In finance, that distinction matters because FactSet and Google Cloud are not only describing research assistance, but planned agents for portfolio operations, deal advisory and corporate finance.
Navigating agent governance and visibility
Google Cloud’s own documentation describes Gemini Enterprise Agent Platform as a system for building, deploying, governing and optimizing agents. The same documentation lists governance components including Agent Registry, Agent Identity and Agent Gateway, with Agent Gateway described as a policy enforcement point for agent tool calls, authentication and security policies.
Google has also been expanding those controls inside Gemini Enterprise. Release notes dated June 25 say Agent Registry can catalog agents and custom MCP servers, while Agent Gateway can set allow and deny permissions on traffic to those agents or servers. The release notes also list agent observability as generally available from June 24.
Regulatory pressures for securities firms
For broker-dealers and other FINRA-regulated firms, auditability is already tied to supervision and records. FINRA’s 2026 Annual Regulatory Oversight Report says existing rules remain technologically neutral and continue to apply when firms use GenAI, including rules covering supervision, communications, recordkeeping and fair dealing.
The same report links GenAI controls to formal review, governance, testing and monitoring. FINRA says firms may consider storing prompt and output logs, tracking which model version was used and when and validating outputs through human-in-the-loop review.
For FINRA, the regulator says AI agents can act autonomously, exceed intended scope and authority, make multi-step reasoning harder to trace and handle sensitive data in ways that may unintentionally store, disclose or misuse it.
FINRA also lists monitoring agent system access, tracking agent actions and decisions and setting guardrails to limit agent behavior as considerations for firms developing AI agents.
Unresolved boundaries in data integration
The MCP element widens the integration question. The Model Context Protocol specification describes MCP as an open protocol connecting LLM applications to external data sources and tools.
In this partnership, FactSet says its MCP and agent-sharing functions will deepen financial intelligence inside Gemini Enterprise. The announcement does not describe the permission model, data-boundary rules or logging approach for that connection.
The cloud component leaves another boundary unresolved. FactSet said it will add Google Cloud to its existing cloud-provider portfolio to improve reliability, scalability and infrastructure capability. The release does not specify where customer data, prompts, responses, citations or agent logs will be processed, retained or investigated during an incident.
