Franklin Templeton Deploys OpenAI Tools to Build Connected Intelligence for Investment Research
AI assistants · Enterprise AI
Franklin Templeton is using OpenAI tools, including Codex, to develop AI workflows for equity research. The firm integrates multi-agent systems and a semantic layer to surface insights across research, risk, and compliance for portfolio managers.
Franklin Templeton announced it is developing AI-assisted workflows using OpenAI technologies, including Codex, to support its equity research and portfolio management teams. The asset manager reported that its initial attempt connecting models directly to raw data failed to yield needed results because the systems lacked contextual understanding. In response, the organization introduced an agentic architecture and a semantic layer representing how corporate data connects. According to Katrina Dudley, Senior Vice President at Franklin Templeton, the initiative aims to combine continuous intelligence from research, risk, compliance, and trading data with human judgment, moving beyond sequential analyst processes toward governed, production-grade workflows equipped with security reviews and data controls.
Key takeaways
- Initial deployments connecting models directly to databases were ineffective until Franklin Templeton implemented agents and a structured semantic data layer.
- The firm uses tools like Codex in practitioner hackathons to develop workflows that counter recommendation bias and surface risk factors alongside research views.
- Investment AI champions transition experimental prototypes into production with defined security reviews, data controls, and governance guardrails.
Source：OpenAI News Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human.