News & Analysis

driven by the PitchBook Platform
Weekend Analysis

Generative AI holds promise and peril for private market investors

Just as venture capital and private equity firms both seek to back the next disruptors in AI, they themselves will face disruption.

Request access to private equity data

The recent hype around AI had a dose of reality this week when users revealed several flaws in Microsoft‘s new AI Bing search.

Powered by ChatGPT, a jaw-dropping AI tool launched by startup OpenAI in November, the service threatens to upend internet search. But it also has a penchant for generating wrong, and often unnerving, answers. Similar issues were found with Bard, a rival product that Google hastily announced after it was caught napping by Bing and ChatGPT. The debacle wiped $100 billion off parent company Alphabet‘s market value.

The technology has some way to go, but generative AI, which uses deep learning and natural language processing models to produce new data, is still likely to have a big impact not just on internet searches, but on many other industries. Private markets will be no exception.

This article appeared as part of The Weekend Pitch newsletter. Subscribe to the newsletter here.

Just as venture capital and private equity firms both seek to back the next disruptors in AI, they themselves will face disruption. AI is already proving to be an invaluable tool for private market professionals in identifying new opportunities, streamlining back-office functions and expediting dealmaking. At the same time, there are concerns that the adoption of AI will have unintended consequences as it creates redundancy and potentially magnifies systemic biases.

A powerful tool

AI has been used by some VC investors in some way or another for years, helping with deal sourcing, due diligence, accounting and hiring. Among them is early-stage investor SignalFire, which closed $900 million in fresh capital this week. Since launching in 2013, the company has been using an AI engine called Beacon to aid its investing and support its portfolio companies.

Chris Farmer, founder and CEO of SignalFire, explained how his firm has historically used AI to collect data on the companies it targets, understand how they compete with one another and benchmark their performance. But now it is using a kind of generative AI seen in applications like ChatGPT.

“The new large language models are now able to understand a lot of those things,” Farmer said, referring to AI abilities to interpret vast amounts of company information. “So rather than doing a bunch of coarse filters, we can actually just throw [the data] into an AI, and we’re using these language models to understand the connectivity between and the similarities [of companies] quite accurately, with fewer layers of process.”

Farmer went on to explain how his firm has used AI for a variety of back-office functions, such as turning term sheets into long-form legal documents or automating the firm’s outreach efforts and generating micro-targeted marketing content.

SignalFire is by no means alone in its efforts to leverage AI. In 2013, Hong Kong-based Deep Knowledge Venture appointed an AI software called Vital as a board member, so it could advise on the firm’s biotech investments. Similarly, London-based InReach Ventures, which touts itself as “the AI-powered VC firm,” uses machine learning for its investment decisions, while Paris-based Jolt Capital boasts its own propriety AI software called Ninja. Also, EQT Ventures started using a proprietary tool called Motherbrain in 2013 to help it identify promising startups.

A boost for dealmaking

The same trend is being seen in the world of PE investing. Motherbrain has already been rolled out to the rest of the EQT Group to be used for its PE investments, while Blackstone now has a data science team that applies machine learning to its portfolio management and dealmaking.

AI, among other things, has been used by PE firms looking to expedite the meticulous and often tedious tasks associated with the deal process.

Doug Cullen, chief product and strategy officer at Datasite, which offers virtual data room software, explains that acquisitions will often need huge amounts of data to be analyzed in short amounts of time. Any tools that can speed up the process, particularly when there are competitive bids, are critical.

“That’s why AI and NLP-powered tools that help dealmakers automate tasks, reduce human error, and ensure greater regulatory compliance, are gaining interest,” Cullen said in an email.

There can be drawbacks, however. As with every new technology, the rising adoption of AI will inevitably render many jobs at a PE or VC firm obsolete. Administrative roles in particular are most likely to be impacted. Farmer notes that SignalFire already boasts a dramatically leaner, yet equally effective, back office thanks to its use of AI. This kind of streamlined way of operating will almost certainly be replicated elsewhere as AI adoption spreads across the industry.

Lessons to be learned

Another perhaps graver concern is that spread of AI will exacerbate systemic biases—which have already been identified as a problem in the industry before factoring in machines. The extent to which this happens will largely depend on the quality of the data used to train AI systems.

“It’s like any other tool,” Farmer said. “A hammer can be used for good, and a hammer could be used to cause damage.” He added that the pitfall of training AI on historical data that has inherent biases is that it could institutionalize structural inequities.

There have been efforts to mitigate specific issues. For example, SignalFire uses AI mechanisms in its recruiting process in order to elevate diverse candidates by adding weight to candidate attributes that are meritocratic in the purest sense, such as open-source contributions.

Regardless of these efforts, AI bias will likely be a significant hurdle for investors to overcome. AI is only as good as the data it is fed on, and as it grows in scope and ability, there will arguably be more variables for engineers and users to consider.

Even when AI is built with the best intentions—and plenty of investment—the negative results can still be unpredictable and expensive. That’s a lesson that both Microsoft and Google are currently learning and that private markets may yet discover.

Featured image by Drew Sanders/PitchBook News

  • andrew-woodman.jpg
    Written by Andrew Woodman
    Andrew Woodman is PitchBook’s London Bureau Chief and oversees news coverage of Europe and the Middle East. Andrew has been reporting on the private markets since 2012. He was previously an editor with Private Equity International and with the Asian Venture Capital Journal. A Japanese speaker, he spent the best part of a decade in Asia, living and working in both Japan and Hong Kong.
Join the more than 1.5 million industry professionals who get our daily newsletter!

I agree to PitchBook’s privacy policy