Franklin Templeton Rebuilds Equity Research With OpenAI Codex

Franklin Templeton is rebuilding its equity research workflow using OpenAI tools, including the Codex coding platform, after an initial attempt to connect AI models directly to a database produced weak results. Katrina Dudley, senior vice president of Public Market Investments, said adding agents and a semantic layer created a "connected intelligence layer" that merges research, risk, compliance and trading inputs for portfolio managers.

How the system works

The firm's first approach loaded data into a database and pointed AI models at it, but Dudley said the models could not grasp how the data related to itself. Adding agents and a semantic layer -- a structured model of what the data means and how its pieces connect -- changed the output. "We've been able to build what we think of as a connected intelligence layer," she said. "We're bringing the data into the tools, and we're surfacing insights that make a difference."

The design goal is to give portfolio managers a single view that draws on research, risk, compliance and trading, four functions that normally see an investment idea from different angles. "We're able to bring that all in and then use that to surface insights to the people who are making decisions, which are our portfolio managers," Dudley said.

Codex in the loop

Codex, which OpenAI launched on May 16, 2025, as a cloud-based software-engineering agent, runs on the codex-1 model, an o3 variant trained with reinforcement learning on real coding tasks. Franklin Templeton has not said which version it uses, nor disclosed contract value, how many staff are involved or any measured performance gains. The firm has used Codex in hackathons pairing technologists with investors, with internal AI champions moving prototypes into production under security and data controls.

Dudley described the shift as moving from augmenting the research process to reimagining the investment research workflow. "We have moved from augmenting our research process to reimagining the investment research workflow," she said.

Research diversity and risk

The firm wants AI to widen the range of research views rather than narrow it, and to flag investment risks explicitly. "We want to make sure that they're accessing that diversity of research and diversity of opinion," Dudley said. She also said, "We need AI to understand that a great investment idea is one that is well reasoned, but it also has analyzed the risks and makes our portfolio managers aware of the risk."

The longer-term goal is to make the connected layer available to every investment team, from analysts up to the heads of each group, with analysts spending more time in the field talking to outside experts and company executives. Franklin Templeton has offered no productivity figures, leaving open how much the rebuilt workflow changes research output versus how much it changes the shape of the process.

Why it matters

The initiative is a test case for whether generative AI can be useful in asset management, where personalization tools that learn user preferences can narrow the range of research opinions. Franklin Templeton's approach optimizes for breadth of opinion and explicit risk disclosure rather than relevance ranking, while keeping human judgment central to investment decisions.

---

Sources

• finance.biggo.com: Franklin Templeton Taps OpenAI's Codex to Rebuild Equity Research Around a 'Connected Intelligence Layer' — BigGo Finance
• securities.io: Franklin Templeton Develops AI Research Workflows With OpenAI – Securities.io
• blockchain.news: Franklin Templeton integrates OpenAI tools for equity research workflows - Blockchain.News

This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.