From the source
- Franklin Templeton is developing AI workflows, starting with equity research, so portfolio managers can compare research views and examine the risks behind an investment idea.
- Codex is one of several tools its investment teams use to build those workflows, according to the company.
- When the firm first tried AI, it put data in a database and connected models to it, but the models did not understand how the data fit together, so the answers fell short.
- Adding agents and a semantic layer, described as a structured model of what the firm's data means and how it is connected, changed the results.
- Franklin Templeton has used Codex in hackathons, pairing experts with investment practitioners as they build their own tools, and its Investment AI champions work to turn experiments into production-ready workflows with security review, expert support and data controls.
- The firm wants its systems to widen the range of research and opinion rather than narrow it, because tools that learn a user's preferences and surface more of the same could bring a biased perspective into the investment process.
