Third Bridge started out brokering one-on-one calls between clients and experts. Its library now holds hundreds of thousands of expert conversations, and Dan Thomas thinks the main reader of that library is going to be an AI agent. On this episode of Invest with AI, Brett sits down with Dan, Head of Partnerships at Third Bridge, who started at the firm ten years ago as a content analyst on European high yield and distressed credit and has published more than a thousand expert interviews under his own name.
Dan walks through why Third Bridge moved early on MCP and what it took to make a library dating back to 2011 work for a model. Before the rebuild, in his words, "We just jammed all the PDFs into the context window." He and Brett also get into what happens to expert calls if those libraries start to look like beta.
What we get into:
Why Third Bridge decided early to be in the agent path through MCP, and why Dan thinks agents will be "fairly unsentimental" about the front end
How the team rebuilt a transcript library that dates to 2011, breaking each transcript into chunks and tagging metadata down to the individual claim, with stable identifiers like tickers and FIGIs
Why vendors shouldn't burn through a client's tokens "willy-nilly," and how cost is becoming part of the buying decision
How LLMs can now pick up the hedging in an expert call that older NLP pipelines missed
What happens when a fund is running 45 MCPs, and why production backtests still lean on the flat file
Why voice agents are now in production for channel-check style work, and why that shifts the bottleneck to the supply of experts
Whether expert libraries become beta, and why Dan expects that to cut demand for "mediocre proprietary one to one research" while pushing funds toward more senior experts
Chapters (Timestamps)
0:00 Intro
0:44 Nickel Smelters, Ferry Companies and 1,000 Interviews
4:34 Who Writes the Transcripts
7:03 Third Bridge's Early Move on MCP
9:10 Your Next Reader Is an Agent
11:45 Rebuilding a Transcript Library for Agents
16:27 Can an LLM Hear the Hedging?
18:56 Click-Through Rates on Citations Keep Falling
22:22 "We Now Have 45 MCPs"
26:05 How Funds Test an MCP Before Buying
28:10 Backtests Still Want the Flat File
30:41 Experts Taking Calls From AI Interviewers
33:34 There's Only One Expert
40:12 Why PMs and Risk Teams Want Expert Content Now
44:05 Mapping Suppliers and Customers Back to Your Thesis
48:24 Are Expert Libraries Becoming Beta?
Follow Invest with AI on:
Spotify: https://open.spotify.com/show/033xcEEovVViS7hIYwNuGZ
Apple Podcasts: https://podcasts.apple.com/us/podcast/invest-with-ai/id1896918892