Private equity is the sleeping giant of the gen AI boom

Ask a room full of people who the biggest winners will be in gen AI and everybody will say tech companies like NVIDIA, OpenAI and Microsoft. A few savvy people might point out that Accenture and McKinsey are printing money. And nobody will mention private equity. Even in a room full of PE folks.

‍

Despite this, PE is uniquely positioned to capture the value that will be created by gen AI over the next decade.

Existing Portfolios

Most PE owned companies are good businesses that will be made better by gen AI. These benefits will come from productivity gains for individual employees, new AI product features that create value for customers, and large workflow automation projects. Not every company will experience all of these, but most companies will find ways to move the needle. 

‍

“AI first” startups taking aim at these companies pose a threat. But incumbents have significant distribution and data advantages that should tip the scales in their favor. 

‍

For example, picture an AI first healthcare startup that aims to automate the process of appealing insurance denials. The incumbent can build the same automations using the same foundation models as the startup (or they can acquire one of the dozens of other small startups going after the same problem). And the incumbent has thousands of input/output pairs of denials/appeals that can be used to train their LLM powered system. They also have deep domain expertise, the ability to start small and automate steps gradually, and a team of experts ready to handle all the tricky edge cases. Meanwhile the startup has to figure out how to get the data they need to automate the workflow, and recreate everything else required for a useful end-to-end solution. And even if the startup somehow succeeds at building a slightly better product, they have to steal the incumbent’s deep relationships with healthcare providers to inflict any damage. 

‍

Unlike in past technology shifts, there are very few inherent network effects creating moats for AI-driven businesses, so startups moving quickly still face long odds compared to incumbents.  Some startups will win despite these odds, but overall it’s a good time to be the incumbent. 

‍

With existing portfolios well positioned, smart firms will turn their attention to a new gen AI investing strategy.

The AI Transformation Thesis

Over the coming years PE will become very good at buying the businesses positioned to benefit the most from generative AI. Here are the primary attributes for an AI transformation thesis:

• Low margins. The lower the initial margins the higher the potential returns. Improving margins from 10% to 20% doubles your profit. Improving margins from 50% to 60% increases your profit by 20%.
• Significant automation opportunities unlocked by gen AI. To find these automation opportunities look for repetitive paperwork tasks and offshore teams running highly manual playbooks. 
• Ability to capture cost savings from automation. Examine the revenue model and pricing power of the business. Consider competitive dynamics including switching costs and the potential threat of new entrants with counter-positioning. Ideally being “first to automate” provides a sustainable advantage.
• Operations bottlenecking revenue. If revenue is bandwidth constrained, then gen AI automations may unlock growth in addition to cost savings.
• Roll-up potential. Once you automate your way to better margins you’ll want to grow quickly through financial arbitrage.
• Margin of safety. Find traditionally good PE bets that are likely to succeed even without a well executed AI transformation.

‍

And here’s the playbook for a successful AI transformation after the business is acquired:

• Align the team. Executing an AI transformation is a huge undertaking. Make it the #1 priority for the company and get buy-in across the entire organization.
• Retain domain knowledge. Employees with detailed understandings of customers, workflows and operations have a huge role to play.
• Hire or partner with A+ technical talent. You need a strong CTO to drive change and talented AI engineers building the automations. The CTO needs to block and tackle to get the AI engineers fast access to necessary data, systems and domain knowledge. While PE firms aren't known for their technical hiring prowess, they can bridge the gap through tuck-in acquisitions and tapping professional services firms with talented AI developers.
• Aim for production ASAP. Start with a narrowly scoped automation and get it into production fast. Then expand to new automations. Do not waste time and money on heavyweight company wide AI infrastructure and data readiness projects.
  • Good: invest in creating high quality datasets of input/output pairs that will be used to train AI systems to do specific tasks. 
  • Bad: invest in central infrastructure that will capture all of your company’s data so that you can use it for everything (“everything” in this context is a synonym for “I don’t know if there are any real use cases”).
• Make intelligent UI decisions. Most automations will involve humans-in-the-loop. Avoid chatbots with steep learning curves and uphill adoption battles. Build automations directly into systems and workflows such that the automated path becomes the new default.

‍

The firms that move fast and execute this thesis well will enjoy extraordinary returns. They’ll be among the biggest winners in the generative AI boom.

‍

Chris Taylor is the CEO of Fractional AI, where he has personally scoped generative AI projects with hundreds of companies. Before launching Fractional AI, Chris was CEO of Xip, CRO of Wove, and held sales leadership positions at LiveRamp.