A whole pig rotates on a spit over an open fire. The Philippine delicacy, lechon, is being roasted for software engineers at Inriver, a data company built for commerce.
It’s a juicy prize for winning a generative AI coding competition hosted by Inriver’s owners, private equity firm Thomas H. Lee. It’s one way the Boston-based buyout firm has been able to implement generative AI coding assistants among its portfolio of middle-market companies.
“As you own a portfolio of companies, you have to be smart about how you introduce new capabilities,” Mark Benaquista, managing director of THL’s Strategic Resources Group, told Business Insider. SRG, made up of 20 people, is responsible for working alongside the portfolio’s leadership teams to provide operational and technology support.
“Sometimes you have to do this, but if we go in under cost cutting, you’re going to get resistance. If you go in under empowerment, you’re going to get receptivity,” Benaquista said.
THL
Across Wall Street, PE firms are eager to realize gains from genAI, which promises to make workers more productive and save time. THL decided to see for itself what impact the technology could have on its portfolio, comprised of companies in the tech, financial services, and healthcare industries. After introducing Github Copilot, an AI coding assistant that helps write and suggest code, across a dozen portfolio companies earlier this year, THL found that engineers, on the whole, were between 10% and 30% more productive with the copilot than doing the same tasks manually.
Engineering teams used the coding assistant for various tasks, such as creating code, documenting it, translating it between languages, or educating others on what a certain chunk of code did. The firm recorded the time it took with the AI versus doing it manually and made a baseline assumption on how much each task represented in the engineers’ overall workload.
Encouraged by the results among software developers, THL is looking to integrate generative AI across its portfolio companies, from sales and marketing to customer service, Benaquista said. Though PE has made a reputation for itself for cutting costs, sometimes at the expense of jobs, Benaquista said that’s the opposite of what THL is after.
The goal is to accelerate software development so that portfolio companies can do more work with the same amount of resources. Put simply: “If your team’s 20% more productive, that, all else equal, should mean you’re getting to 20% more of your product roadmap,” Alex Sabel, a research analyst at THL overseeing the business aspects of SRG’s work, told BI.
“Just because this is becoming easier to do doesn’t mean you’re going to need less software developers,” Sabel said. “It actually could possibly mean you need more software developers in the future,” because the pace of development could mean applications could be revamped every month or week, instead of every few years, he said.
On the ground: convincing skeptics, competitions, and a pig party
Flying around the globe to help get portfolio companies up to speed with the tech is Jagjit Singh, a director within SRG with an engineering background. He’s racked up thousands of miles flying between his Boston home base and India, the Netherlands, and the Philippines. His two tactics for getting skeptical engineers on board? Gamification and a little tough love.
“If you don’t adapt it, it’s here. So the bear is already here, closing an eye doesn’t mean it goes away,” is one of the messages Singh tells tech teams. He also persists that deploying generative AI coding assistants is not a cost exercise, but a means to productivity. No developers have been let go in this exercise to reduce cost, he added.
A little friendly competition with alluring incentives helps, too. Singh often pits two engineering teams against each other to see who can win the prize, or “badge of honor.” Knowing the audience and what’s relevant to them is key to dialing up the motivation. Recent prizes have included the latest Apple Watch, an Apple TV, or even — as mentioned above — a whole roasted pig party, Singh said.
“It’s a pride, it’s not about the money. It’s about presenting a lechon to somebody. It’s a big deal,” Singh said of the pork specialty.
Coding assistants are just the beginning for THL. Already, the firm is ruminating on other ways it can jumpstart business with AI, like shifting through unruly data for teams responsible for sales and pricing, Sabel said. There are other opportunities within customer-service and internal operations teams, Benaquista added.
With portfolio companies realizing early wins themselves, the pipeline for more AI tools is growing long. Now, the key for the SRG is educating portfolio companies on what problems AI can fix and which it can’t. We don’t want to just throw it at any problem, Benaquista said.
“This isn’t free,” Benaquista said. “Our cost structure in some cases has gone up. We believe it’ll flow through into greater productivity and show it’ll show itself somewhere else.”