Private equity (PE) firms are seizing the benefits of artificial intelligence (AI), which we define as computer systems that can sense their environment, then think, learn and take action in response. In our recent survey of US companies actively using AI, nearly a quarter of respondents came from PE-backed companies. Here are the top benefits they see, the top obstacles they report and — based on the experience of PwC experts working with PE-backed firms on AI initiatives — ideas for how to overcome those obstacles.
PE portfolio companies are ahead of the pack with AI. A full 31% (compared with 23% of companies that aren’t PE-backed) are AI leaders: They report widespread adoption of processes fully enabled by AI. It only stands to reason that companies backed by PE would be so advanced with AI, which offers a powerful set of tools to unlock value. Wisely, PE portfolio companies are using AI for far more than just automation to cut costs. Their top goals for AI initiatives include increased productivity, stronger product and service innovation, and revenue growth.
A PE firm seeking to unlock value from a manufacturer faces a challenge: how to estimate the likely demand for the company’s various products in order to optimize production. It’s a complex question and not only because of always-present macroeconomic uncertainties. Given the wave of recent economic upheaval, many of this hypothetical company’s consumers and business customers have changed their consumption habits. To forecast demand, the company needs to know to what extent these changes will persist — as well as how its competitors and suppliers will respond. Pre-AI, the most the company could do was extrapolate a range of possible forecasts based on historical data, then map out the implications for demand.
With AI, this company can build a digital twin to create a far more sophisticated, accurate and dynamic model. The AI-powered digital twin accounts not just for different market forecasts, but how the company’s own production choices might impact the market; how its competitors might respond; and how its workforce and supply chain will hold up. As new data comes in through APIs or website scraping, the twin ingests it and keeps updating forecasts so the company can keep monitoring its production and optimizing revenue growth. To fill gaps in historical data, this twin has leading AI that synthesizes data based on probabilities and alternate data sources.
Many PE-backed firms are already using AI with success (see box for example). At the same time, others are getting stuck with their initiatives — often in the same places.
PE firms tend to be highly skilled at achieving economies of scale in their portfolio companies. That’s critical for AI: Since AI technology and talent can be so expensive, you really need to scale it to achieve a satisfactory ROI. But to scale AI, you need AI models and data sets that can be used across the company. Unfortunately, silos between lines of business and functions often make that hard.
Besides, another big bottleneck for AI initiatives is talent. AI specialists (such as data scientists, machine learning engineers and model operations engineers) are in high demand. It can also be challenging to upskill your non-tech experts to take full advantage of AI tools. Culture is often part of this problem: A lot of company executives don’t feel comfortable with AI. It’s hard to make a business case for AI — which may require a significant tech upgrade — to people who don’t even want to use it.
A PE firm is looking to improve sales at a recently acquired retail chain — and to do so, they need to understand exactly what is driving (or failing to drive) sales today. This hypothetical retail company has both online sales and hundreds of brick-and-mortar stores, divided into several brands. Across the brands, the firm has identified several dozen stores that have static revenue despite favorable locations. For deeper insight, the company turned to AI. It plans to use cameras and IoT sensors in stores (with full attention to privacy rights and regulatory compliance) to precisely measure sales of different products as they relate to foot traffic, time of day, weather and recent advertising campaigns, among other variables. They will then plug this data into a digital twin, which will analyze all the data and simulate the impact on sales and profits of changing up store inventory or altering marketing.
But this company’s different brands have different data systems; some individual brands even have different data systems for different product lines; and the company’s suppliers have their own data systems. Until the company can collect and standardize all this data, their simulation may not be complete or accurate.
Fortunately, it’s possible to overcome these obstacles — with the right focus. For companies that are already deploying AI and for those that are behind the curve and need to catch up, the following five guidelines can help achieve big benefits faster.
Choose AI that does its own data work. AI runs on data: up-to-date, in scope, standardized, verified and accessible. The best AI solutions help build their own foundation with tools to ingest, reconcile, verify and standardize data from multiple sources in real time.
Start with costs. To start a virtuous cycle of AI investing, prioritize cost takeouts. Opportunities for cost savings from AI are common, especially among portfolio companies’ legacy digital assets. They can produce funds to pay for further investments in AI.
Go to the cloud. Building on-premise technology stacks for AI is often too costly for portfolio companies. The cloud can be both cheaper and faster. If you choose the right cloud provider and carefully attend to the shared responsibility of data ownership and governance, security and compliance risks can be managed.
Make it responsible. Roll out better governance, transparency, security and compliance so you and all your stakeholders can trust your AI. To do that, you need to build not just an AI solution, but an AI asset, which includes checks and balances and quality control on your AI.
Build out capabilities. Some initial AI deployments may be plug and play, especially when done through the cloud. But as solutions advance, many portfolio companies will need to upskill their workforce. They’ll also need to build a culture and a structure that can take advantage of greater automation and quickly act on real-time, data-driven insights.
PE firms are justly famous for transforming companies they acquire. With the right approach, AI can help them plan, execute and monitor that transformation, unlocking value every step of the way.