Skip navigation
Advisors and AI
david-plon-promo.png Image created using Midjourney

The Minds Behind the Machines: David Plon, Portrait Analytics

Portrait will be able to answer any question or perform the typical tasks one would ask of a junior analyst at a hedge fund, including finding investment ideas, building financial models and creating pitch decks.
Resources

At Portrait Analytics, David Plon and his five-person team are building an AI-powered “junior analyst.”

Plon, the startup’s co-founder and CEO, knows the role well—he was one for several years with The Baupost Group in Boston, paying his dues pulling together investment reports from company filings and reports.

While Plon knows the workflow and processes and the data that make up an analyst’s daily work, Connell Gough, his co-founder, has the technology chops as a full-stack developer and heads up the engineering team.

The goal: Portrait will be able to answer any question or perform the typical tasks one would ask of a junior analyst at a hedge fund, including finding investment ideas, building financial models and creating pitch decks.

“Down the road, we hope and envision creating something that anyone involved in investing can use—I’d love, in five years, that my mother’s own financial advisor had access and used this on a daily basis,” said Plon.

The initial product is a question-and-answer application for users to query the system about the public markets.

Portrait was founded in 2022 but only exited stealth mode in early April with the announcement of a $3 million funding round led by .406 Ventures alongside a few hedge funds.

Since launch, the startup has rounded out its engineering team and started providing access to some early pilot users and investors, though the client list is still under wraps, he said.

“It’s still very early days but we have been working on lots of features we hope to complete in the coming months and have started to integrate SEC filings, 8-Ks, 10-Ks and tables,” Plon said, in addition to other forms of publicly available regulatory filings and searchable information like earnings’ call transcripts.

Eventually, the system will let users incorporate their own proprietary data as well, siloed off from other users.

For now, the tool is wading into the available data to produce summaries and reports for users, who in turn are helping Plon and the team shape the final product. “They are folks from within our pilot customers, but a large enough group to get good feedback signals,” Plon said.

Over the coming months Plon will roll out access to more analysts on the company’s waiting list. “What we are attempting to create is a system that is both useful and reliable, and there is not one answer as to how to get there,” he said.

He likens the process to creating a factory with machines in a kind of assembly line producing a final unified product. In some cases, the team can use simple open-source models as a starting point, then do a lot of bespoke development to create something unique.

“Some of what we are doing is leveraging some of the larger foundational [learning] models”; other things the team must build themselves.

“The key thing I’d say is much like a human analyst, you don’t lose trust if you say I don’t know, and only provide answers where you are super secure in what you are saying and can reference back to a source piece of content—you cannot give users false information or analysis,” said Plon.

Next: Harald Collet, Alkymi

TAGS: People
Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish