We often talk about companies as if they are people, but really, it’s the leaders behind top firms who are the real Agents of Change. That’s what inspired this series, which gets to know the investors and thinkers pushing funds to think of themselves as innovators and early adopters, especially in a market where a slight edge can mean everything.

For every wave of technical innovation, there are the innovators who absorb the risk of experimenting with new technologies on behalf of the entire market. It’s often not until a second segment of early adopters help bridge the gap to a broader audience that mass adoption is seen. But by then, many firms, both on the tech side, and the customer side, have already missed an opportunity to be early movers, define the tech in a way that suits them, and get a jump start on capturing its gains.

As we spoke with over 100 investors in private equity, we realized it’s often PE associates who are at the forefront, navigating the complexities of investments and being pushed to think about doing things faster, better and smarter.

Traditionally, the Associate role is one of the most sought-after destinations for MBAs, junior financial analysts and consultants alike. Their days are filled with financial modeling, due diligence, and direct negotiations – vital tasks that require both deep analytical skills and a nuanced understanding of human relationships – with an almost immediate impact on real businesses.

How Is AI Impacting the Associate Role?

Today, investor tasks are being augmented by AI, introducing both opportunities and concerns. We have observed that many funds are pulling together innovation teams, with associates often making up the bulk of the group, to think about where to apply new technologies in a way that actually has a practical business impact.

In our conversations, we also heard about what we view as less ideal approaches. Some funds are jumping straight into partnership discussions with big names like Microsoft and OpenAI to provide tools like ChatGPT Enterprise or Microsoft Co-Pilot to employees. Others are siloing innovation to a select group of teams and evaluating the impact of new point solutions on a small sample.

While these strategies can help, we agree with approaches recommended by AI experts like Ethan Mollick, who started his “flipped learning” model in the classroom, and now recommends similar distributed models to enterprises. In these approaches, employees would be given greater resources and latitude to experiment with AI on their jobs, then come back to their teams to solidify and disperse learned best practices.

This also aligns with the “Decentralize” strategy recommendation from the last discussion in our Products & Markets series. In the context of the investing world, where AI use cases are taking shape but often still unrefined, it helps to have the practitioners closest to the subject matter provide input on AI application. Often fresh out of business school or with several years experience at a top investment bank, associates have seen a breadth of scenarios, and have the depth of knowledge to evaluate impact on the ground.

To be sure, we decided to interview a handful of associates in greater depth, and kick off our Agents of Change series with perspectives from these powerhouses of PE deal teams.

Structure Remains, With Room to Explore

Last month, we interviewed four associates across middle market private equity firms, a fund at a major US bank, a small internationally focused firm, and a global mega-fund. While all agreed to share their insights, most also requested to remain anonymous in order to be able to speak freely about their experiences and their firm's respective strategies.

Experiences can differ vastly for associates at diverse PE firms, but one common theme we observed is a desire to think and experiment beyond what most funds are doing today. Analysts and associates are still limited in their actual ability to experiment with technology, often due to firm-wide policies, or simply a lack of resources. Many firms we spoke with outright block access to AI tools like ChatGPT. Those who do allow access to LLM-based sites emphasized that they be used primarily for initial research and simple text summarization, never modeling or final outputs.

This makes sense given the sensitive nature of the data and the state of most AI-powered point applications today. Because the risk of a deal going south can be outsized, firms maintain rigorous standard operating cadences to reduce the likelihood of an undesirable outcome, whether from.

Still, other funds are being smart about structured experimentation. While a narrow approach to testing AI technology may not be as powerful as decentralized adoption, it’s better than nothing at all. Other associates we spoke with mentioned that innovation and testing new approaches is encouraged, but the market is simply not ready to deliver solutions that drive meaningful impact quite yet. Once we dove deeper, it became clear what deal teams need to see before going all-in on AI.

Pain Points and Solutions

While often described as more manageable than investment banking, workstreams owned by PE associates are still done on a very fast turn-around, especially at smaller funds or those participating as minority investors. “Just about every deal cycle is the same. Our diligence process is very rushed. Typically there is a deadline to submit bids in less than two weeks, and we're just hearing about it on the Friday before”, described one associate we spoke with. To help with this, several ideas surfaced in our conversations:

Dashboards: A common theme across all of the requested products we spoke about was enabling greater real-time insights into deal analysis and portfolio operations. “A dashboard, with data refreshed on a weekly or more regular basis, to allow AI to pull in information about sales or operations. Putting it in a format where we can see how the company is performing in real time,” would be a game-changer, per a Senior PE Associate we spoke with.

Modeling Assistants: Pilar de la Barra, at the Latam-focused Australis Partners, talked about the potential to increase efficiencies in Excel-based work. “I'm still at a point where I review almost every cell in the model, so figuring out how to efficiently go over some of the technical analysis, that's a rather tedious task that I could see being improved.”

Modeling, of course, continues to occupy much of the time of deal teams, with work often spread across VPs, associates and analysts. Though some IT and operations associates talk about the role of AI in analyzing large datasets and rapid scenario modeling, the reality is that little of this is powered by generative AI today.

3rd-Party Data Integrations: While most investors have access to premium market intelligence, and capacity to acquire more on a per-deal basis, the process of bridging the right data into the due diligence and modeling process can still be time consuming, and leaves associates wondering if the data used to supplement their assumptions is the best of what is available.

Data coverage creates an opportunity as well, with market data in developed markets being much more readily available than in emerging economies. Still, De la Barra’s involvement in operations and board decisions at Australis' portfolio companies underscores a trend in which firms are increasingly curious about technological tools to streamline operations and enhance decision-making. As emerging markets in Latin America continue to grow, with Colombia's PE market expanding by over 15% in the last year alone, the integration of sophisticated tools could become essential in remaining competitive with North American firms.

Barriers to Adoption

While some investment associates facing these pain points are eagerly pushing for solutions, many are also more realistic and measured in their optimism for AI. Perhaps due to the incentive structure in private equity, even entry level investors feel a strong degree of ownership over their work, and are highly tuned into the concerns of central management related to feasibility, compliance and operations.

Privacy and data integrity are major examples of such concerns, where even associates who would not be directly affected by data leakage or IP issues are aware that deal information must be handled very securely.

More relevant to the analysis itself, investors are typically among the two-thirds of knowledge workers aware of issues with hallucination and inaccurate data in LLM training data. “We are very wary of where we are getting our data from”, mentioned an associate at a major bank.

Associates at internationally focused funds described working on the front lines of deciphering data in different formats and diverse operating norms across different markets. Nonetheless, they are often not quite ready to trust AI solutions in without significant human oversight.

The result? One associate tells us, “honestly, we're still pretty old school. We use AI in basic tasks like writing internal memos or summarizing news, but today, not much more.”

Taking the First Step

The operational involvement of PE associates varies significantly between firms, influenced heavily by the size and focus of the fund. Some roles are more narrowly defined, focusing on financial analysis without direct operational control, which is typical in larger funds where investments are more about financial structuring rather than operational turnaround. Here, AI can serve as a critical tool in managing complex global portfolios, offering real-time insights that were previously unattainable.

In theory, these larger firms could see faster gains from automation as work is already more clearly defined and operationalized. For instance, AI tools are being leveraged to perform rapid due diligence, a task that some noted involves "stress tests" on financial models provided by potential investment targets. Such tools not only speed up the process but also enhance the accuracy of projections and risk assessments.

Associates we spoke with at smaller firms often talked about wearing many hats, and, in some cases even participating in conversations at the firm leadership level, which might be too complex to benefit from AI just yet, or at least require more custom-tailored applications.

For now, it is unclear whether AI will serve to further entrench economies of scale at large funds, or help make smaller funds more competitive. Perhaps the answer is both.

Nonetheless, a top debate in AI continues to be whether the technology will ultimately add value and create new roles by enabling novel capabilities, or simply make processes more efficient and possibly slow hiring for knowledge work roles.

As you might expect, associates talented enough to make it into leading PE funds do not see their work as being immediately replaceable by AI, nor should they. When investors look at the technology, those we spoke with see opportunities to automate away some of the busy work or become more competitive in their roles, rather than see entire steps of the process disappear. As one associate from a large cap fund told us, “at the end of the day, PE associates are energized by their roles and want to do them better, not less often."