Q&A: Value creation: the impact of AI on private equity
October 2024 | SPECIAL REPORT: PRIVATE EQUITY
Financier Worldwide Magazine
October 2024 Issue
FW discusses value creation and the impact of AI on private equity with Alberto Fumo, Jan Timmermann, Sumit Malhotra, Ben T. Smith IV and Evan Gutoff at Kearney.
FW: Could you provide an overview of how artificial intelligence (AI) has penetrated the private equity (PE) landscape in recent years?
Smith: Artificial intelligence (AI) has permeated several aspects of the private equity (PE) industry to varying degrees. AI tools are being deployed to automate operational processes and increase value for the firm and its clients. AI is also a target for investments. The potential in AI for PE is clearly there. Today’s opportunity is in optimising PE firms’ internal processes to ultimately add value to portfolio companies. For example, a fund recently launched an integrated generative AI (GenAI) programme for its employees across the globe. One of its use cases is to use GenAI to summarise data from lengthy documents typically received by limited partners (LPs). Executives are also now deploying GenAI to identify subsectors for investment, request for proposals writing, coding, report generation and to analyse contractual language. Another large global PE fund invested in an AI-driven portfolio monitoring and valuation solution. These tools cannot replace the human judgment of investment principals, but they can assist them to go deeper faster across more targets.
Timmermann: The use of AI tools and technology for automation and creating value has been increasing since the success of ChatGPT and other large language models. Concurrently, the coronavirus (COVID-19) pandemic has accelerated openness to AI adoption, in the way it has for other technologies. Firms started to use AI for everyday tasks such as data management, outreach efforts and back-office processes. However, there is scepticism about these tools given the lack of clear return on investment (ROI). Funds want AI providers to offer more clarity in workflows and outputs, practical use cases, and how to ensure data security risks are properly managed. They need to be sure that proprietary pre-deal or portfolio company data does not inadvertently make it into the public domain. There is also a feeling of general mistrust of nascent technology. There may be a first-mover advantage as there always is, but that advantage is not yet clear.
Fumo: According to a 2024 Private Funds chief executive survey, about 50 percent of funds are exploring potential use cases for AI implementation. However, only a small proportion of funds surveyed have implemented the technology. We have some way to go before AI tools are commonplace in PE firms’ operations and working practices. According to Pitchbook, there was 23 percent growth in PE-led dealmaking between 2019 and 2022 in AI and machine learning (ML) technology companies. This was influenced by the post-pandemic era in which openness to new technology meant higher rates of adoption, coupled with higher applicability and availability of attractive investment opportunities. According to Pitchbook, China, the UK and the US have led investment activity in AI and account for over two-thirds of deals in the past five years.
Malhotra: While there has been activity in dealmaking, the pace has slowed since 2022. Funds are cautious and are assessing ROI, especially in implementing new technology. This has been affected by other, broader economic factors such as the high inflation environment, rising interest rates, economic deceleration because of these factors, high valuations, and ongoing, globally impacting geopolitical tensions. Firms will continue to scrutinise investment and buyout opportunities in technology companies building AI-ML tools and understanding the impact of AI across front, middle and back office operations, while being cautious when it comes to deals to maximise returns.
Gutoff: A compelling investment opportunity today for PE funds is in technology infrastructure that supports AI. Investments in data centres and energy assets have garnered attention. Robust underlying infrastructure is critical to broad AI adoption, regardless of industry sector, and will only increase as AI becomes more powerful and widespread. PE firms are already beginning to invest in energy and infrastructure, positioning themselves to benefit regardless of which tools and use cases win in the short term.
“A compelling investment opportunity today for PE funds is in technology infrastructure that supports AI. Investments in data centres and energy assets have garnered attention.”
FW: What are the driving forces behind the PE industry’s adoption of AI? What value creation opportunities does it present?
Malhotra: There are several triggers for the adoption of AI. The adoption of technology has generally accelerated since the pandemic, and GenAI applications have proven themselves to be successful more broadly. The inherent urge to gain a competitive advantage by leveraging new technology has also driven adoption. We are still in the early phases of embracing GenAI for PE and most firms are waiting to see evidenced and quantitatively positive impact of AI-ML technology. That will take time, given deal cycles and the nature of how PE deal opportunities are currently identified. At the same time, there are some examples that show progress. For example, Schroders Capital has created an AI platform called GAiiA initially rolled out for its PE practice. GAiiA expedites the analysis of large volumes of data, so that Schroders’ PE investment specialists can focus on strategy.
Gutoff: There are many opportunities for AI to deliver value creation for PE, both in the PE fund operations and the operations of portfolio companies. On the deal side, AI could increase efficiency in deal sourcing and analysis by identifying companies, leveraging large amounts of data at an extraordinary pace. AI can also help drive PE firms’ efficiencies, such as support in report generation and investor communication strategies, customer relationship management, data analytics, marketing strategies, legal work and general business processes. The largest impact, however, will be how AI is leveraged to impact portfolio companies. Funds that can harness the power of AI to optimise revenue management, selling, general and administrative (SG&A) costs, and other key business drivers will have a major edge in generating returns and being able to bid competitively for select assets.
Timmermann: At the same time, there are many hurdles before these efficiencies can be achieved, such as data quality and accuracy, data privacy concerns, and the complexity of integrating new technology with existing client systems that could limit widespread adoption. AI models also need to be tailored to analyse and understand the many types of businesses that comprise PE fund portfolios. Our understanding is that proprietary, in-house solutions are currently more successful than third-party applications. Two large, global US-headquartered alternatives fund managers have led the charge so far, leveraging AI to analyse market trends, evaluate potential investments and enhance decision-making processes.
Fumo: Many PE portfolio representatives do remain sceptical, with some parties feeling they have not yet seen anything game changing from AI. Likewise, there are too many risks associated with AI, particularly with respect to multibillion-dollar deals. In such deal transactions, parties do not delegate tasks to AI. These concerns can be seen as a strong indication that there are clear risks that PE firms are alert to in automating the use of AI.
Smith: There are obvious opportunities to leverage AI to find deals. However, there will be an explosion of other capabilities that will enable deeper due diligence for investors. For example, access to data owned by the target asset will glean deeper analysis of potential deals. We can see this shift in the old-line image and text media companies where licences fundamentally change the economics of a deal. Even $10m of 100 percent margin data licence revenue changes a deal dramatically. The technology to support GenAI is changing the value of everything from access to energy to real estate. That is not only in relation to data centres, but also about access to power – for example a power contract held by an old factory that is monetised by that contract being sold to GenAI data centre operators.
“The goals for GenAI should align with the corporate and strategic priorities of the firm, whether those encompass cost reductions, growth or anything else.”
FW: Could you outline the range of AI applications for PE? How would you characterise its transformative potential?
Smith: Due to the highly competitive environment for new deals, PE firms face a laborious task in their first stage of originating deals in analysing potential targets. They must identify opportunities early from reams of data, assessing it efficiently so they can build conviction for a deal fast. With the support of an in-house data science team, or through partnerships with AI companies, the latter could provide an intermediary approach to AI for PE. Depending on the strategy or the information to be managed, PE firms are developing customised deal targeting platforms to power predictive analytics by scanning proprietary databases, filtered by requirements driven by the investment team, and fed by a broad range of data sources, from investor data to market and sector data, to news and social media content. The automation of that process could free up investment teams to focus on the deals they want to pursue, rather than spending time assessing reams of data to find those deals.
Malhotra: There are some good examples in the market today. A large European fund has applied AI to investment targeting. It has built a platform that handles millions of data points and selects potential investments with a high likelihood of success, driven by prompts from the investment team. The tool was first used by venture capital (VC) and has now been rolled out for other asset classes. The firm has also launched a cross-functional working group of almost 400 individuals who are testing and innovating in the use of GenAI. Among the use cases is the ability to summarise data from long documents, including those from LPs.
Fumo: The most advanced use of AI by PE is in the portfolio management space. Portfolio optimisation tools can help to improve efficiency, drive higher profitability, identify new sources of revenue and reduce risk. For example, a large global PE firm is currently using AI to automate monitoring and benchmarking across its portfolio of companies from transaction data to operational results and forecasts. AI can also have a role in a fund’s exit strategy since that process is well-defined and managed by the financial sponsor. In exits, data metrics can be standardised and stored, which means the potential for AI solutions is high. Natural language processing can be used to extract data from a multitude of unstructured documents, such as financial statements and subscription files, or to automate processes such as capital calls or distribution of capital statements.
Gutoff: However great any solution sounds, there are barriers to implementing GenAI tools. One is data availability. Many companies have tried to address this challenge and continue to do so. We also need to understand the readiness of investments for AI technology. The tech stack of companies during due diligence is key to understanding how much of the AI playbook for value creation can be deployed for a specific company. Ultimately, as we have seen for innovative technologies in the past, investors will seek companies relevant in the new, AI-driven world, so it is imperative that funds and their targets are ahead.
“Funds are cautious and are assessing ROI, especially in implementing new technology.”
FW: Drilling down, in what ways can AI technologies help PE firms to identify and realise value opportunities across their portfolios?
Malhotra: There is significant potential for AI in portfolio monitoring, since there is a huge volume of unstructured data that cannot be manually analysed, such as the management of inventories or working capital. AI solutions can be used for existing processes to improve efficiency and profitability. We expect tech-forward companies will streamline and improve business processes and improve customer experiences, for example in marketing and sales departments to generate more leads. They will integrate AI in the product itself and transform its customer-offering, as well as improve efficiency across the value chain, for example by lowering the cost of serving customers or helping to compete in a market when the company has resources in many geographies to operate more efficiently.
Gutoff: AI solutions are also being used to make existing software as a service (SaaS) based solutions more automated, streamlined and powerful. AI solutions can also be leveraged to create a higher degree of personalisation based on user behaviour, increasing its power further.
Fumo: In deal identification and screening there have been some advancements in use among VC funds. Some funds are training models on what they are looking for so the AI tool can seek out newly founded companies that meet their criteria in terms of investment area, founder background and so on. There are tools being developed for financial analysis too, that can analyse a market accurately and achieve that much faster than humans, which will be a great first step, and an excellent way to increase junior analyst productivity.
Smith: An incubator called ‘8090’ provides inspiring food for thought. It is built based on the insight that you can re-engineer existing SaaS products at a fraction of the cost because engineering costs have reduced so much, given the more advanced productivity of AI. That means existing products will be replicated at a fraction of the cost – 80 percent of the features for 10 percent of the cost is the impetus for the name of the incubator.
FW: What regulatory considerations do PE firms need to make when utilising AI technology? How might compliance issues affect AI deployment strategies?
Timmermann: Regulatory uncertainty is one of the main risks of investing in AI. All countries and regions are considering how best to control AI and its use without stifling innovation. Initial frameworks for these regulations are currently being written, so firms may be conservative when it comes to investing till there is more visibility. One of the major concerns is the threat against data protection and privacy. The risk of losing data is a major concern for PE firms. Before evaluating and deploying GenAI platforms, firms must ensure that data will not leak, at the minimum. Some experts are saying that a central challenge is the fact that everything is moving so quickly. An OpenAI solution does not have the operational security or functionality available with, for example, office software, so it is not yet fully professionalised.
Fumo: There are some interesting case studies. A PE fund based in Scandinavia established an ethical compliance forum targeting the use of AI before launching its pilot programme. The forum defines guidelines and the scope for AI use. For example, inputting personal data such as performance reviews is forbidden, which is essential from an EU General Data Protection Regulation compliance perspective. The same approach was taken for client data. This highlights how important it is to have rules on what can be inputted and how outputs can be used. This tool does not allow the use of firm-specific information in OpenAI’s ChatGPT. The firm plans to develop a detailed AI policy governing what the technology can and cannot be used for, outside of imminent regulatory guidelines.
Gutoff: We believe there are six categories of risk when it comes to developing or deploying GenAI. And firms can mitigate these risks through a thorough assessment process that looks at the elements of each risk. The first is cost of ownership. The cost of deploying GenAI should be considered in terms of accessibility, then scalability to ensure the strategy works long term. The second is data risks. Data risks can encompass inaccurate or biased model outputs due to insufficient or poor-quality data. The third is AI model risks. If there are inaccurate or inexplicable outputs from a specific model, there may be model risk in that GenAI tool. The fourth is privacy and security risks. GenAI can pose a significant data security risk which can impact an individual’s privacy and related cyber attacks. The fifth is risks to copyright and intellectual property (IP). There is a host of challenges within the risk to copyright and IP as GenAI use cases can breach copyright and IP laws with unauthorised content generation embedded in tools leading to legal implications. The sixth and final risk category is ethical and social risks. GenAI use cases create the potential for deepfakes, misinformation and misuse of individuals’ personal information.
Malhotra: Ultimately, the key to maximising the potential that automation of GenAI tools can offer requires human intelligence and interventions to integrate with unbiased and ethically sound data, with a deep understanding of industry specifics. GenAI is not in itself ethical or unethical; its actions are dictated by biases in model design, underlying algorithms and the training database. There are also risks in other areas; any service business will be at risk of being outcompeted by GenAI services just as any SaaS incumbent is at risk of an AI challenger shifting the competitive landscape completely. There is huge potential for disruption across many business models.
Smith: For PE firms to effectively mitigate risk, they will need to do some challenging groundwork by breeding a culture of agility, responsibility, transparency and decision making that is scalable. Firms will need to rethink their operating models with a focus on AI as a prerequisite to implementing and using new technology. And the significant amounts of money at stake coupled with nascent GenAI governance can exacerbate the risks. The time to mitigate those risks is at the outset. There is also a big picture view when it comes to risk that is important to note. The national security implications of GenAI today are at an equivalent level to the security risks of the Manhattan Project back in the 20th century. This creates complicated closing risks with Committee on Foreign Investment in the United States and other regulatory frameworks that are being extended into new areas.
“All countries and regions are considering how best to control AI and its use without stifling innovation.”
FW: What essential advice would you offer to PE firms looking to expand their development of AI technology? What steps can they take to mitigate potential risks and maximise benefits?
Smith: Given the early stage of development of GenAI for PE, there are a host of variables that firms should evaluate before deploying the technology at scale. The goals for GenAI should align with the corporate and strategic priorities of the firm, whether those encompass cost reductions, growth or anything else. Use cases should focus on tangible value and be defined by clear metrics for success, as well as the right operating model and market. Firms should focus on identifying existing synergies, rather than reinventing the wheel. The key to achieving this is to understand what the firm is already doing and integrate GenAI with those strategies, focusing on its value-adds while accounting for additional risks. This is akin to what happened in 1995 during the early cycles of investment in the internet. Investors must consider the threats that GenAI presents to an investment, the secondary opportunities for new types of companies it creates, and the natural hype cycles.
Timmermann: A key prerequisite for successful deployment of GenAI is quality of data and standardisation of metrics. The penetration of new solutions in PE is slower than for other industries as funds rely on metrics and data supplied by their target businesses or from portfolio companies which are rarely consistently structured. This pre-work to prepare data requires significant time and resources, such as data teams, warehouses and so on, but it is valuable since firms will gain a better understanding of data using less technical methods before encountering more complex technologies such as AI models and ML. We do not yet know but it may be that standardisation of metrics for PE only works at the very highest level. If you are looking across, for example, 160 private businesses covering multiple industries for a given fund, the finer details of those businesses will be too specific to standardise for GenAI.
Fumo: The devil is always in the detail. Data on targets such as revenue per full-time employee, cost of goods sold as a percentage of revenue, SG&A as a percentage of revenue, percentage of earnings before interest, taxes, depreciation and amortisation, net working capital ratios which funds already receive – all these metrics that contribute to key performance indicators vary enormously within and between industries. Whether these data can be standardised remains to be seen. It will depend on each fund.
Malhotra: Risk assessment is of the utmost importance. PE firms should make sure they invest in the right skills and knowledge to circumvent potential future risks of any GenAI technology deployment. Ultimately, the best solutions are those that are easy to integrate, flexible, scalable and secure.
Gutoff: PE firms looking to expand their development of AI technology need to get their data strategy right and begin running experiments immediately.
“Portfolio optimisation tools can help to improve efficiency, drive higher profitability, identify new sources of revenue and reduce risk.”
FW: What AI-related trends do you expect to shape the PE industry in the years ahead? Is this technology set to revolutionise long-established practices?
Fumo: As PE firms expand their use of GenAI and the technology becomes more sophisticated, we are likely to see increasing levels of integration into routine operations. Primarily, deal sourcing and market diligence platforms could become more robust and increase efficiency gains to reduce the lead time for PE funds’ preliminary processes. But it is early days, and we are yet to see if firms will use externally built platforms or develop their own to achieve these efficiencies.
Timmermann: More refined AI systems could expedite the analysis of large volumes of unstructured and structured data, drastically reducing lead time, improving data accuracy, and removing human error and bias, which would help maximise and isolate returns-driven decision making.
Smith: There are some interesting developments ahead. AI and GenAI technology could evolve from solving back office automation processes to helping automate large scale enterprise platforms. We could see AI assisting in identifying optimal exit timings and strategies by analysing investor sentiment, market conditions and performance parameters. The more data is available, the more sophisticated these predictions could become, helping investors to maximise returns.
Alberto Fumo is a senior partner at Kearney and leads the firm’s global private equity (PE) and principal investors practice. He has spent over 20 years advising corporations and PE firms across the full transaction and transformation cycle. He specialises in assessing acquisitions and divestitures from multiple commercial and operational perspectives and supporting management in the post-acquisition phase. He can be contacted on +44 (0)20 7468 8862 or by email: alberto.fumo@kearney.com.
Jan Timmerman has over 20 years’ experience in industry and professional services. He has managed over 50 PE assignments across sectors including business services, consumer products, energy, healthcare, life sciences and technology. His skill set covers operational due diligence to post deal-transformations, including post-merger integration, transformational carve outs, operating model design, operational restructuring, strategy development and implementation, and finance transformation. He can be contacted on +44 (0)20 7468 8441 or by email: jan.timmermann@kearney.com.
Sumit Malhotra has been on the inside of multiple mergers, acquisitions and carve outs. In a career spanning more than 23 years, he leads Kearney’s PE, technology due diligence and post deal value creation business across Europe. One specialty he has cultivated is maximising the value of portfolio companies through technology and digital. His client work focuses on pre-deal IT due diligence, large-scale technology-driven transformations, and performance improvement programmes, including cost optimisation and synergy assessments. He can be contacted on +44 (0)20 7468 8721 or by email: sumit.malhotra@kearney.com.
Ben T. Smith IV, one of the firm’s youngest partners, is a board member and global lead for Kearney’s digital and analytics practice. He is a key adviser to Kearney’s large clients and its board of digital transformations, strategy, acquisitions and investment. He supports the development of the firm’s venture capital relationships, where he founded Kearney Venture Capital to drive innovative and impactful acquisitions. He can be contacted on +1 (415) 490 3675 or by email: ben.smith@kearney.com.
Evan Gutoff heads the firm’s foresight and innovation team, which is comprised of a global network of tech pioneers, founders and investors. Through this team, he aims to help executives unlock new opportunities at the intersection of technologies, business models and start-up ecosystems, while bridging the divide between stagnation and acceleration when it comes to innovation-driven growth. He can be contacted on +1 (617) 415 5472 or by email: evan.gutoff@kearney.com.
© Financier Worldwide
Q&A: Value creation: the impact of AI on private equity
Achieving a successful exit through an auction process in today’s market
Looking beyond leverage to drive sponsor returns
Recent trends in the German private equity market
From crisis to confidence: leveraging alternative assets and data for investment security