The deal sourcing process promises to get a lot faster and a lot better, thanks to generative AI. PE Hub spoke with a range of dealmakers at private equity and folks at third-party origination services to find out what’s possible today and what’s coming tomorrow.
With the ability to pull a significant amount of data about potential companies to invest in and then draft up an email to a prospective company that a firm wants to invest in, generative AI is quickly evolving into a regular piece of the investment process.
Already, deal sourcing service providers like SourceScrub and Finquest are using a combination of traditional AI with generative AI. Traditional AI represents web scraping and data processing, while generative AI involves training a learning model to produce original content such as music, art or words.
“Where AI comes in is: it helps scale, helps show recency and it’s comprehensive,” SourceScrub CEO Jim Obsitnik told PE Hub. “We have our proprietary data set that we’ve been cultivating. But then there is this large language model that came along with ChatGPT that represents the collective knowledge of the internet all the way up through 2021. We’ve created the interconnection of those two data sets.”


SourceScrub, based in San Francisco, matches raw data and derived data points and provides an investor lens for firms that use its platform, added Obsitnik.
Finquest’s goal is to make the data that comes from generative AI digestible for firms, Tanguy Lesselin, co-founder and CEO of the Singapore-headquartered company, said.
Though there’s been a significant amount of talk around ChatGPT, Lesselin maintains that generative AI as a whole is only as good as the underlying data it is trained on.
“There will eventually be a proliferation of niche applications that leverage those algorithms,” he added. “The winners are going to be [ones] who are going to have data, and that can make that data more valuable to users.”
Interest in generative AI is still in the exploratory phase, with some PE firms beginning to experiment with it. Traditional AI is more common, and still an evolution from decades ago.



In the past, it would take a PE firm a week or several weeks to pull lists and add contacts. Next would be a process of inputting the contacts into a customer relationship management system and then creating a mail merge.
“If a source servicing provider can boil the ocean for me and identify 100 targets that we feel really good about and I can get emails out to those people in 24 or 48 hours versus the olden days, the vault, the volume and the velocity that these tools can bring to deal sourcing is tremendously valuable,” Jessica Ginsberg, managing director of business development at lower mid-market private equity firm LFM Capital, told PE Hub.
Deal flow impact
Traditional AI has made a significant impact on deal sourcing, speeding up what was once considered an arduous process.



“Having both AI and data-driven models for both mergers and acquisitions and deal sourcing strategies is critical to pinpoint relevant investments and advisory opportunities,” Dominic Chan, director and head of financial sponsors group for Vaquero Capital, told PE Hub.
Traditional AI gives firms the opportunity to know who to reach out to among the large swath of technology companies. It also leads to higher conversion and success rates by streamlining the due diligence process.
Chan said since using software like SourceScrub, the deal flow pipeline has increased 40 to 50 percent.
Some companies, including KPMG, have already started to integrate generative AI into their own data platforms. Chan said it can help sift through research, earnings, transcripts and pitch decks. It can also extract and consolidate a lot of key points for briefings.
As Vaquero collects more data in the research process, it creates its own proprietary models where it scores what opportunities and companies are the best targets to spend time on or travel to and go visit for a pitch.
Drawbacks
Though generative AI is having its moment in the sun, there still are questions to address.
There are sometimes “hallucinations” within generative AI, warns Josh Giglio, vice-president of product at SourceScrub. Unless a conversational model, such as ChatGPT, has been pre-prompted, it may respond with inaccuracies in the interest of completing and continuing the conversation.



Another challenge is blending a private equity firm’s staff with automated processes without causing disruption.
“I’ve got a team of people that are building relationships with business owners,” Ginsberg said. “For me, what’s front of mind is, does it make sense to use something like [generative AI]? And then can we still leverage the team that we built and have them work in tandem?”
It’s clear that generative AI can help PE firms discover potential leads, but it can’t be the one making the decision.
Generative AI will help level up the whole industry in numerous ways such as building familiarity in different industries and sectors, Lesselin said. The limitation is that humans still must be the ones making the decisions, which may sometimes come down to gut feelings.
“There’s no forward thinking [with generative AI],” he added. “It’s trained on the past. You can only be as good as what the past has been. And that’s a major limitation, because when you’re an investor, you’re looking at the future, not at the past.”