With the explosion of AI applications, private equity houses and their portfolio companies must understand where key opportunities lie.

By Tom Evans, Kem Ihenacho, David Walker, Laura Holden, Hector Sants, Claudia Sousa, Catherine Campbell, and Patricia Kelly

Click for larger image.

Artificial intelligence (AI) developments provide increasing opportunities for private equity, including deal sourcing and portfolio company analysis/enhancement, particularly in businesses that can adopt a customer subscription model or leverage big data opportunities. However, the adoption of AI technologies, and investments in new AI businesses, pose significant challenges. To ensure that time and capital are deployed productively, firms must understand the market space and usage for these tools, and the workings and accuracy of any underlying technology. How technology models and algorithms work, where underlying IP resides, and where data is stored are key. Whilst the use of AI is often discussed, it is much less often understood; we are seeing an explosion of AI applications and PE houses and their portfolio companies need to understand where the opportunities are for them to exploit.

A Tool to Secure Deal Opportunities and Drive Portfolio Company Growth 

According to a survey conducted by Intertrust, 90% of private equity firms expect AI to have a transformative impact on the industry. AI-backed data analytics are playing a growing role in analysing and identifying deals. QuantCube Technology, for example, provides in-depth data analysis, drawing on customer reviews and social media posts to develop predictive indicators of events, such as economic growth or price changes. There are now companies offering AI-driven technologies that claim to help source PE deals. While this presents a potentially compelling use of AI for investors, it remains to be seen whether these technologies will deliver results. 

AI technologies used to speed up legal due diligence can be applied to commercial and financial diligence and other aspects of the deal process, bringing time and cost efficiencies to investments. Further, AI has portfolio company applications — from the financial sector to consumer and retail, AI is driving back-office efficiencies in HR, IT support, cybersecurity, and data aggregation, resulting in cost savings and quicker decisionmaking. AI can also improve front-office functions, and is increasingly used to analyse and predict customer trends.

However, the introduction of AI tools widens the scope for unexpected outcomes. Firms and portfolio companies must understand what products actually do in practice, a task that can be difficult when software is not proven or is self-learning — developers may not yet fully understand capabilities and can be hesitant to stand behind guarantees. Firms should consider how their business will contract for this technology. Whether by acquisition, licence, joint venture or otherwise, each method carries specific risks. A key question to consider is who owns the models and the resulting algorithms. As systems are made “smarter” by training, firms must understand if learning is shared, and what the commercial and legal implications could be.

Consider the Investment Opportunity

According to a report from KPMG, a total of US$12.4 billion has been invested in AI technologies to date, with a predicted US$232 billion in deal flow by 2025. For private equity firms looking to invest in companies offering AI-based technology, issues that hamper commercialisation and licencing of software will be critical. AI businesses are essentially software businesses, so concerns around the creation and ownership of intellectual property matter. Having clarity on who has developed the code and in what capacity is critical. Further, acquirers of AI tools targeting financial institutions need to be cognizant of end-user restrictions on security and data storage for cloud-based AI tools. Firms must also understand the application and commercial impact of AI ethics regulation.

Data Feast Brings Risks and Challenges

AI algorithms improve themselves by feasting on data. The more data is consumed, the better AI tools become at spotting patterns, such as customer behaviour patterns that train a bank’s AI systems to better spot fraud. However, the risks of using AI technology underpinned by huge data sets cannot be underplayed, with maximum fines under the EU General Data Protection Regulation (GDPR) now being the higher of €20 million, or 4% of an entity’s global turnover. The cost of potential liabilities for breach/misuse can exceed the purchase price on a deal, and if issues are identified, the value of an acquisition will be reduced, as GDPR fixes are expensive to action. Deal teams must carefully scrutinise whether AI businesses are compliant, or risk assuming significant liabilities.