Start-up FinChat aims to disrupt research space using GenAI

Braden Dennis, CEO and founder of stock research platform FinChat, does not have a background in finance. An environmental engineer by trade, he last worked as an environmental supervisor at Ontario Power Generation, an organization owned by the Ontario government in his native Canada.

Over a call in early May from FinChat’s Toronto office, Dennis explains that he nurtured his adolescent interest in financial services by starting a podcast, the eventual success of which encouraged him to pursue finance full time.

“On my 18th birthday, I started investing,” Dennis says, “I’m 28 now, so I’ve been investing for over 10 years. During that time, I started a podcast from my dorm room to learn from the authors and investors that I wanted to learn from. No one listened to that podcast for like six years—like, no one. But in 2019, it took off.”

During the pandemic, Dennis began working on Stratosphere, an investment research analytics platform—and FinChat’s predecessor—as a product for the audience he had gained off the back of his podcast, The Canadian Investor. At the time, he did not foresee the tools he was creating being in particularly high demand from non-listeners. He was wrong.

We recognized that [generative AI] would be a new way to interact with data, so it was more of an experiment than anything. Like, what happens if you rethink how we can query all this data? That was the genesis of FinChat

Braden Dennis, FinChat

“I started working on Stratosphere as a product for that audience, thinking that it would just be something small,” Dennis says. “I was charging nine bucks a month for it. When I got it to, like, hundreds of customers, I said to myself, ‘Hey, screw it, I’m going to go full time on this.’”

Staying humble

Dennis and his team had been working on Stratosphere for two years when ChatGPT launched. Like many others, he saw an opportunity for experimentation.

“We recognized that [generative AI] would be a new way to interact with data, so it was more of an experiment than anything. Like, what happens if you rethink how we can query all this data? That was the genesis of FinChat in the spring of last year,” Dennis says.

FinChat, a subscription-based public equity research platform, uses a ChatGPT-esque chatbot interface that allows retail investors and portfolio managers to access aggregated reports on a wide range of public companies. AI is used to locate, aggregate, and summarize reports into manageable and easy-to-parse sentences, and users can use tools within FinChat to create models and visualizations, and query individual call transcripts and earnings histories.

Its launch on April 20 of last year garnered a lot of attention. While it had taken Stratosphere more than two years to hit 50,000 active users, FinChat reached 100,000 in a month. Dennis, not wanting to be caught up in what might be fleeting, remained wary that by launching a product in the middle of a hype cycle, his company was potentially attracting the wrong kind of interest.

“It was useful to have that kind of virality out of the gate, but they were the GenAI users, not the buy-side users,” he says. “We didn’t get too cocky, because we knew a lot of people wouldn’t be serious customers, so we had this huge launch and then a trickle-down. It didn’t go to zero, but it reached a base level, and from there we’ve been growing really nicely.”

FinChat and Stratosphere merged last November, combining the data research capabilities of Dennis’s first company with the AI tools of his second. Dennis says FinChat’s investment research assistant, also named FinChat, queries institutional data in a way that separates it from just being ChatGPT under a proprietary guise. He says this distinction is illustrated by how the product deals with requests about non-US companies.

“The way we use the language model is like using a translator to go fetch data that we know is accurate and correct, instead of it searching for some crappy internet article, or parsing through a filing and confidently pulling the wrong number out,” he says. “Parsing is useful and good, but it’s terrible once you leave the US, and we’re bringing global coverage to this.”

Dennis explains that while working on Stratosphere, the company acquired a large amount of useful datasets that were categorized and aggregated in-house, and FinChat leverages this data to complete tasks. Dennis’s team was frustrated that nobody was licensing out some specific metrics that investors look for, such as non-GAAP earnings, so they began collecting the data themselves.

The speed at which these younger, newer organizations can turn around updates is phenomenal. It separates them from some of the bigger companies, and these tools are only likely to improve productivity, and most importantly, our—the buy-siders’—productivity

Portfolio manager at an equity fund

“We thought it was ridiculous that you could find over 35 years of financial statements, but you couldn’t find out Netflix’s revenue or membership count,” he says. “I would have to go through and find those numbers myself, so we started aggregating that data in-house. When we found this new way of querying data [with AI] it was a new product, and when you looked at the industry of tools that were available, we were all of a sudden first to something.”

Proof in the pudding

Behind Dennis’s desk is a large table packed with people, deep into their work. It appears that this is the norm for FinChat, especially when Dennis explains how the team celebrated the product’s first anniversary in April.

“The anniversary was a message in Slack, and then back to work,” he says.

Dennis’s efforts to expand and improve FinChat are reaping some rewards. As an early mover in generative AI, and with a product to show for it, his company is in a vaunted position, having secured $1.5 million from US-based fintech investor Social Leverage in December of last year.

But being an early adopter in a new field comes with its pitfalls.

“Look at everything that was the hot thing in the late ’90s,” he says. “It wasn’t usually the first thing that won, so that’s what I always tell my team. Just because we were first doesn’t mean we should be cocky or think we have this in the bag.”

Like other companies, such as BlackRock and Microsoft, FinChat features a copilot, an AI-powered assistive tool that can aid users in their workflows. Dennis says he’s bearish on copilots’ place or utility across all companies and verticals, but explains that in market intelligence, accurate summarization abilities are crucial.

“Most investors, especially if they’re covering a wide universe, don’t have the capacity to read, listen to, or ingest millions of data points,” he says. “I’m extremely bullish on [copilots] being useful for this vertical because search didn’t work for this industry. CTRL+F didn’t work for this industry. Files are scattered globally, there’s no structure for any of them, and search doesn’t work well for the combination of structured and unstructured qualitative [data] and structured, numeric quantitative data together.”

A portfolio manager at an equity fund who uses FinChat says its smaller size allows it to be more agile in rolling out updates that use new technologies than its larger, more lumbering competitors.

“I think the speed at which these younger, newer organizations can turn around updates is phenomenal,” the portfolio manager says. “It separates them from some of the bigger companies, and these tools are only likely to improve productivity, and most importantly, our—the buy-siders’—productivity.”

The market intelligence space is dominated by two large players, namely Bloomberg and S&P Global, with the rest of the market split among several smaller, boutique market data providers, each with their own niche. One of FinChat’s value-adds is its focus on individual companies’ key performance indicator (KPI) data, a carryover from Stratosphere, which was launched for listeners of Dennis’ podcast with the main goal of aggregating public companies’ KPI data.

A portfolio manager at a family office says that one of the advantages of sourcing investment data from FinChat is that due to the company’s size, it’s easier to get in touch with the team when problems or queries come up, and because of FinChat’s subscription price, it’s easier for smaller buy-side firms to afford the data it provides.

“Braden’s team is very responsive,” the family office portfolio manager says. “With other companies I’ve used, it can take weeks to do something, but for their customer service, it is the same day. I had a much more expensive platform that I used, and when that contract ran out, I didn’t renew and just went with FinChat full time. It’s 90% cheaper and has all the functionality I need.”