Franklin Templeton Builds AI Research Workflows With OpenAI

OpenAI has published a customer story describing how Franklin Templeton, the asset manager, is using its technology to support investment research. The entry point is deliberately narrow: equity research comes first.

According to the case, Franklin Templeton is developing AI workflows that help portfolio managers compare research views and examine investment risks. Both tasks sit in the information-integration and risk-checking part of the research process rather than in drafting or client communication.

The scoping language matters. Starting with equity research implies room to extend into other asset classes or stages of the workflow, but the public material covers only the equity research piece for now.

Research is a natural landing zone for AI in asset management: the material is dense, cross-comparison is routine, and conclusions need to be traceable. Placing the model at intermediate steps such as comparing views and checking risk also leaves the final judgement with human portfolio managers.

The case does not disclose which model is used, how widely it is deployed, how many analysts or managers are covered, or any quantified gain in speed or accuracy. What can be confirmed is the direction of the partnership and the purpose of the workflow, not validated investment performance.

Compliance context is the other thing to watch. Investment research involves large volumes of non-public material and documentation requirements, so once an AI workflow enters this layer, audit trails, data boundaries and the explainability of model output become part of the purchasing decision.

The next signals are whether the workflow expands from equity research into fixed income, multi-asset or advisory work, and whether Franklin Templeton publishes usage scale or outcome data. For peers, the reference point is not which model was chosen but how far the research process itself has been rebuilt.