A year into the explosive advent of generative AI, it’s become increasingly obvious that these technologies truly do promise game-changing disruption for many industries and business functions.

While there’s been no shortage of wild-eyed exuberance since ChatGPT launched in late 2022 and showed the world what generative AI is capable of, there’s been no shortage of substance, either.

Generative AI is a critical reasoning engine capable of having an open-ended conversation with a customer, producing rich marketing content, and scanning vast stores of data to provide deeper insights. The impact across industries and business functions is already plain (see Figure 1).

In private capital, we see firms mobilizing in three important ways:

• Within the portfolio, they are wasting no time gaining insight into the potential impacts and opportunities across companies and industries. Will these technologies disrupt a portfolio company’s value chain or economic model? Is there a chance to lead the change by leveraging generative AI? With a short list of companies in hand, firms are producing results by linking initiatives to strategy and executing through rapid test-and-learn efforts. They’re finding that focus and change management matter—a lot.
• In due diligence, leading teams are developing scorecard-based protocols to assess generative AI threats and opportunities in every diligence the firm takes on. The aim is to make it as routine as legal or commercial diligence. Firms are also using AI tools to speed up and sharpen the underwriting process. In many cases, generative AI presents the unique opportunity to build prototypes in diligence that can rapidly prove (or disprove) a disruption thesis.
• At the firm level, generative AI offers plenty of ways to streamline or automate back-office functions. But its real power is to dramatically expand the scope of information that the firm brings to bear on investment decisions. These tools can leverage a firm’s scale by making institutional knowledge instantly available to everyone who needs it.

Drawing a bead on value in the portfolio

Like any technology, generative AI is best deployed as a tool in service of strategy. It doesn’t create value by itself but by linking explicitly to pragmatic business objectives. How can we better serve our customer? Which metrics are we trying to inflect, which processes are we trying to improve, or which people are we trying to make more efficient?

Scattershot initiatives will not drop any money to the bottom line, but a series of use cases targeted at a specific role or process very well might. While starting now with a test-and-learn mindset is critical, it’s as important to prioritize investment against the initiatives most likely to deliver the highest value.

For CVC, that meant applying a generative AI lens to more than 120 of its portfolio companies across geographies and investment strategies. Starting from an industry perspective and drilling down to the company level, the firm asked several key questions for each asset: Is the underlying customer need likely to change? Is the business model under threat? Will generative AI enable new competitors? And what barriers to entry or competitive moats exist to protect against disruption?

CVC sorted companies into three buckets: those that face revolution in the very short term, those whose business is likely to transform over the next few years, and those where meaningful disruption is unlikely. This analysis helped the firm prioritize which companies would benefit most from investment, and when. 

One of them was Italian online educator Multiversity Group. Education is widely viewed as a sector with potential generative AI applications, and CVC saw that Multiversity was uniquely positioned to develop a strong AI-enabled business model. The company had robust market share, fully accredited content, and participation in a highly regulated university landscape. Even before generative AI came along, it was developing a set of initiatives to improve everything from how students enrolled in classes to how they interacted with professors. 

The million-dollar question was how the company could accelerate those efforts using generative AI. The answer was setting up an “MVP accelerator” to identify generative AI applications, develop the business case, build a minimum viable product (MVP), and test and learn to refine a solution.

One example was using generative AI modules to answer routine questions from students about class content or administrative issues that take an inordinate amount of a professor’s time. The initiative removed 80% of those questions from professors’ plates, allowing them to redistribute that time to more value-added activities like course planning and one-on-one interactions with students. This benefit to instructors helped ensure their adoption of the technology.

The MVP accelerator put as many as 30 initiatives in motion and institutionalized the company’s ability to innovate. It not only buttressed Multiversity against competitive incursion but will also burnish the company’s exit story. Going through this exercise at Multiversity and other companies in its portfolio, meantime, has turned into a master class in generative AI for CVC. Scanning the portfolio is making the firm smarter and enabling it to be more responsive when it comes to deploying these technologies.

Bolstering due diligence

Many of the questions around threat and opportunity in a portfolio scan are also foundational to the most effective generative AI diligence scorecards. Additionally, it’s important to assess organizational readiness. Has the company developed a vision for how it can deploy these technologies? Does it have a data strategy, and has it developed use cases? Is the right talent in place to execute, and does the company have a track record when it comes to innovation?

As generative AI gains speed, it will become increasingly critical for firms to institutionalize this kind of scrutiny. Deal teams should be doing a fast analysis of any target company, asking whether generative AI is likely to have an impact—positive or negative—in the years ahead. The quick answer may be no. But anything other than that is worth investigating further.

For one firm targeting an IT services company, that meant determining whether generative AI could make key functions more efficient. Given the company’s rapid growth, the objective wasn’t to cut staff. But the potential new owner wanted to project whether future growth could be made more profitable using AI.

An analysis of various knowledge-work tasks across the company suggested that several departments could do more with less by automating certain activities and using AI to speed up others. That could generate margin improvement of 10% to 15% in the midterm as revenue expanded—enough to give the buyer an added layer of conviction that the target would be able to justify its multiple.

Due diligence teams can also use generative AI to get a more complete picture of a target company’s prospects. Powerful tools are rapidly emerging to scan reams of data in a fraction of the time it would take a human to do the same job. One Bain & Company tool can ingest 10,000 customer reviews, print charts, and summarize findings within minutes. Another can summarize interviews with customers and market participants, converting unstructured text data into structured formats. These tools widen the aperture to more information, more quickly, so deal teams can focus on generating insights and testing their investment theses. Generative AI helps them pinpoint the market research and competitive analysis needed to underwrite specific opportunities.

Unlike many technologies, generative AI also lends itself to building live models in diligence that can help “prove disruption.” This came into play recently when one large buyout firm was considering the acquisition of a company that had built a proprietary AI-based tool for a highly technical industry vertical.

The tool was designed to transcribe and summarize data to create efficiency in the customer’s workflow. The company had trained it extensively on proprietary data, and the selling point was that it could process this complex technical information with a standard of accuracy critical to the company’s customers.

Very quickly, however, the diligence team demonstrated that the tool faced a serious threat in the marketplace. In a matter of days, the team built a series of prototypes using OpenAI’s GPT-4 API and other open-source models. They then tested these “competitors” against the target’s solution and found that all of them performed significantly better in a number of ways. This allowed the fund to quickly make a call on the opportunity.

Sharpening fund insights

What’s becoming increasingly clear to general and limited partners alike is that generative AI can have a transformative effect on all manner of fund operations. Many firms are looking at how AI tools can take out costs in the back office and otherwise make internal operations more efficient. But more and more are thinking strategically: How can we use generative AI to supercharge our investment professionals across the full value-creation cycle, from sourcing, screening, and diligence through to portfolio management and exit?

The goal here isn’t to fill seats with less expensive robo investors but to make investment professionals smarter and faster at what they do. One large investor at the forefront of thinking through these issues is backing generative AI initiatives that cut across the investment cycle. The most advanced is a project to help investment professionals become more productive by speeding up (and improving) the bread-and-butter busywork that is critical to sourcing and evaluating deals. The firm also hopes to vastly expand their access to information and sharpen insights about both target companies and the macro conditions in which they operate.

Generative AI tools are ideal for scanning massive pools of data for insights using the firm’s preferred screening criteria. In deal sourcing, this can be invaluable. Currently, this fund’s professionals tend to look at 10 deals to find 1 worth investigating further. Armed with a set of seven key criteria linked to the fund’s strategy, they spend a full day on most “looks,” or half a day if they’re lucky. Generative AI can not only help produce the initial list faster but can also bring down the screening time per company from a day to an hour. This makes team members significantly more productive and frees them up to focus on the more qualitative work involved in analyzing the potential gems that make it through the funnel.

Gaining conviction around those potential winners also benefits from generative AI. Chatbots can help investment professionals easily leverage the firm’s scale by sorting through its prior internal work on the subsector or market. They can crawl through more data—both internal and external—than deal teams could ever do on their own. Consider that, for most midsize target companies, you can’t find a published source of Net Promoter Score^SM data or any other objective measure of customer loyalty. Scanning a selection of customer reviews might get you somewhere. But AI tools allow you to scrape every review ever posted to the Internet within minutes, organize the comments meaningfully, and then generate a clear, analytical report. You end up smarter than you would be otherwise in a fraction of the time.

Getting started

For funds asking themselves how to start using generative AI to improve investments and investing, now is the time to get educated on potential impacts and begin planning for them. Here’s a practical priority list:

• Even if you do nothing else, scanning your portfolio is critical. Generative AI is moving fast enough that you need to understand today which companies might face major disruption and which may be able to ride that disruption to better performance.
• With a short list of companies in hand, link AI initiatives to clear strategic objectives that can boost performance, like customer satisfaction, revenue generation, or cost reduction. Centering on a specific product, process, or group of workers will have much more impact at the bottom line than a diffuse approach leading to small gains that are difficult to bank.
• Don’t neglect change management. Clear governance and a detailed execution plan are essential to making change that sticks. People are nervous about AI. If you don’t make it obvious how the technology will improve jobs (rather than eliminate them), frontline employees will refuse to embrace the technology or actively undermine it.
• Bring what you are learning to sourcing and due diligence, and start building fluency in the generative AI tools that can make you smarter. It’s obviously critical to avoid walking into disruption, but you’re missing an opportunity if you aren’t thinking strategically about how generative AI can drive better returns in your area of focus.
• Begin thinking about how generative AI can transform your own shop by giving investment professionals a leg up on the competition. It’s still early days, but now’s the time to start adopting the AI-enabled tools and processes that can corral data and information in powerful new ways.

As with any disruption, generative AI is both unnerving and exhilarating. But it pays to get moving. Zeroing in on the handful of companies and initiatives that will create real impact is the best way to start turning these disruptive technologies to your advantage.