The potential and pitfalls of AI for private equity

Taking a measured approach to evaluating AI tools arms PE firms to navigate the opportunities and risks of various potential solutions.

By Gianluca Rossi, Ontra

Heightened regulatory scrutiny and increased pressure on fees and expenses have pushed private equity firms to improve operational efficiency. Firms are considering digital solutions to optimize investor onboarding, fund management and dealmaking. Most recently, artificial intelligence (AI) has gained traction, promising to significantly improve legal processes.

As a provider of AI-enabled solutions tailored to the needs of private markets, we believe that AI offers tremendous opportunities. However, firms must understand both the potential benefits and pitfalls of AI.

One application of AI to the private markets that has garnered significant attention is the potential to improve non-disclosure agreement (NDA) processes. For PE firms, NDAs continue to be a point of frustration. Many still use manual and outdated processes that negatively affect their business. In fact, according to a recent Ontra survey, 58% of respondents said their firm’s traditional NDA process had hurt their ability to close a deal in the past year.

To accelerate the process, firms might consider generic generative AI solutions to draft and review NDAs. In some cases, it may offer a good starting point for generating appropriate clauses within a prescribed set of rules. However, AI alone is still not a panacea.

Humans in the loop

Today’s AI tools are often trained on broad data sets not specially tailored to the private markets. They can introduce irrelevant or off-market clauses, omit key language, propose unnecessary revisions or ignore deal context. These errors can prolong negotiations and create friction with counterparties, ultimately diminishing desired efficiency gains and cost savings.

By contrast, an AI solution that includes “humans in the loop” may provide a better solution. Pairing AI tools with human legal expertise can allow firms to streamline legal drafting and negotiation through a hybrid process, in which AI offers suggestions and lawyers exercise professional judgment.

By analyzing and making inferences from thousands of documents, AI can accelerate the negotiation of high-volume, routine documents and surface relevant precedent. Lawyers can then leverage these insights to complement their professional judgment, working faster and more efficiently, with a precision that humans alone cannot achieve. In our experience, this hybrid approach has shown the potential to accelerate negotiations by 50% and decrease costs by nearly 70%.

Post-negotiation, firms can also use AI to improve ongoing contract compliance, such as monitoring obligations included in side letters to LPs . For example, AI trained on relevant data sets is adept at generating clause summaries, extracting key dates, and highlighting key covenants. Lawyers can audit AI output and operationalize improved compliance processes. Firms can quickly streamline cumbersome and error-prone manual processes, allowing professionals to focus on generating returns for investors.

Deciding where and how to apply AI tools to legal processes requires a thoughtful understanding of mission-critical workflows. Pressures to streamline are real, but firms should be prudent in assessing and adopting AI. Particularly for recurring high-volume and time-sensitive workflows, like negotiating NDAs, there is no margin for error. Implementing the wrong solution may cause disruptions that result in worse outcomes than manual processes and take considerable time to unwind.

Here are five tips to help you consider AI solutions for your firm:

Prioritize outcomes and stakeholders. Even routine contracts can directly impact your business, risks, relationships, and reputation, whether the process is cumbersome or efficient.

Clearly define use cases (e.g., contract types or workflows) and determine the desired outcomes.

Research and evaluate AI capabilities to ensure they align with your specific requirements and are suited to private markets.

Test AI tools before relying on them. Verify accuracy and quality by comparing outcomes to existing processes and required standards.

Consider what implementing AI tools will require from you. Understand whether outputs will require extensive review from you, or whether outputs will be vetted by humans before they come to you. Human expertise and review are crucial for ensuring AI outputs are legally sound, compliant, and aligned with industry practices. It is important to understand how and when that review will occur and who will be expected to perform it.

AI is nascent and still developing. Taking a measured approach to evaluating AI tools arms PE firms to navigate the opportunities and risks of various potential solutions. Starting your AI education now to reduce the learning curve for evaluating next-generation tools as they emerge. When approached judiciously, AI can address evergreen pressures to reduce PE operating costs and improve efficiency.

Gianluca Rossi is director of machine learning at Ontra.