Read more
Thought leadership
Insights
Artificial Intelligence and Data: The New Frontier in Private Equity Investing
Thought Leadership
Author: Jan Bruennler, Managing Partner
January 2025
Featured in the 2025 edition of the Evercore State of the Market magazine
While the importance of data and the transformational potential of artificial intelligence (“AI”) is now widely recognised, the commitment and capabilities of mid-market GPs in their use of data varies widely. As a leading software investor in the European mid-market, Bregal Milestone has positioned itself at the forefront of the AI revolution. At the inception of our firm, we took the decision to establish Bregal Milestone as a data-centric investor, aiming to be a leader in the use of AI and data across all aspects of the investment lifecycle. We invest in continually augmenting and improving our proprietary AI platform and using the most advanced technology available. We also benefit from our partnerships with some of Europe’s leading innovators across our portfolio. This has afforded us a unique vantage point and a privileged perspective on the new frontier that has been established through the transformation of the private equity industry by technology.
Before delving into how AI and data informs our approach to investing, it is worth taking a step back and asking: what are the problems we are trying to solve?
Discovery and origination
The first problem relates to discovery and origination. As we know, in an efficient (public) market, the clearing price of an asset is an unbiased estimate of the true value of the asset. In theory, in an efficient market, in our opinion, any actively managed and researched strategy would not outperform a strategy predicated on randomly diversifying across assets.
In intermediated private market auctions, it could be argued that the situation for an asset buyer is worse. An auction creates momentary liquidity to purchase only. The clearing price is the highest bid, with no countervailing selling pressure at a high price, leading to systematic overpricing in certain market environments.
And yet this doesn’t tell the whole story. In the economist Burton Malkiel’s classic story, a professor who espouses Efficient Market Hypothesis is walking along the street with a student. The student spots a $100 bill lying on the ground and stoops to pick it up. “Don’t bother,” says the professor. “If it was really a $100 bill, it wouldn’t be there”.[[1]]
At Bregal Milestone we believe that the $100 bill is there, but it is unlikely that it is just lying there waiting for us to pick it up. It might be located off the beaten track, hidden and may not look much like a $100 bill at all at first glance.
Let’s consider the characteristics which promote efficiency in a private equity market:
- Homogenous and widely available data
- One language and legal system
- Transactions available in bulk viaaggregators (financial advisors)
- Price discovery via auction
- Identifiable cohort of companies, steady over time
In contrast, the European software mid-market which we invest in, looks rather different:
- Highly fragmented, inconsistent and insome cases unavailable data
- Many languages and legal systems
- Many transactions created one-by-one inpartnership discussions
- Price discovery via negotiation
- Dynamic cohort of companies, changing constantly
In other words, we are fortunate to be facing a market which promotes market inefficiency and increases the chances of finding that $100 bill, provided that our ability to pick it up is superior to that of our competitors.
This is a feature rather than a bug of this segment of the market. Within it, we often find highly ambitious entrepreneurs, seeking to stay invested in and involved with their companies, tying their personal wealth creation to the continued development of their company post-transaction. They care about the capabilities and value system of their investment partners and seek to solve highly specific problems in a fragmented and complex European market. Success requires a level of mutual trust and understanding between a business owner and private equity investor, which is often built over years rather than months.
As a result, in the European software midmarket, we track a highly dynamic group of approximately 35,000 companies in 30 countries, with management teams speaking 25 languages. Moreover, due to structural growth and innovation, approximately 1,400 companies enter the cohort every single year.[[2]] Lagging or imprecise data will not deliver an edge in uncovering proprietary investment opportunities in our space.
Data and AI need to support our efforts to discover European software entrepreneurs.
Value creation
Secondly, data and AI can be used to improve our ability to achieve greater creation of enterprise value during our investment horizon through strategic input, operational improvements and capital allocation.
The benefits of thoughtful AI adoption are becoming evident, as highlighted in Figure 1 further on. Organisations that have adopted generative AI report attaining significant cost savings and revenue increases from AI investments in areas such as software engineering, IT, marketing and compliance. However, the same analysis shows that companies often struggle to see a return on investment early on in their AI journey. As of 2023, the majority of respondents still reported no revenue increase and no costsaving through adoption of generative AI.
Figure 1 – The State of AI. [[3]]Organisations most often see meaningful cost reductions from generative AI use in HR and revenue increases in supply chain management
In the mid-market, where resources are more limited than in large organisations, this issue is exacerbated. As a knowledgeable private equity partner with proprietary AI capabilities, we accompany our portfolio companies on this journey and direct AI investments to specific use cases which offer tangible impact and lasting value creation.
As a second requirement, our data and AI strategy therefore need to be able to reliably produce operational improvements and enterprise value creation post-investment.
Data and AI strategy at Bregal Milestone
We are guided by the mission to be “the leading investment firm in our market in the application of data and AI, generating superior risk-adjusted returns for our investors.”
At Bregal Milestone we have developed a proprietary AI platform, Mosaic, over many years and it is the result of this vision. Its proprietary code base builds on our own in-house data engine and a wide range of integrations with relevant data sources.
Today, Mosaic is integrated in every facet of our investment process. It draws on every single data point in our firm and digests data volumes at terabyte-scale to create actionable intelligence for our investment and value creation teams.
If a sector team researches trends and data in a specific software sub-sector, identifies potential investment targets, develops a point of view to discuss with an entrepreneur, forms a view on company valuation, maps and analyses potential tuckin acquisition targets, or dissects a company’s customer data cube to suggest improvements in account management or marketing spend, Mosaic collaborates with team members in each of these activities. Our team is able to complete each step in a fraction of the time that manual intervention would require (in some cases in real time in a meeting) and does so accessing proprietary private data in addition to public data sources.
AI-driven value creation in action
The Bregal Milestone team uses Mosaic to support the organic and inorganic growth of our portfolio companies. One example is Bregal Milestone's portfolio company Allshares, a European leader in compensation and benefits software. The Allshares investment is an example of our thematic approach to software investing, coming off the heels of a highly successful investment in ePassi, in an adjacent space, which grew its EBITDA by 7x over four years under our ownership and was exited in 2023.[[4]]
In 2024, Bregal Milestone secured a majority stake in Allshares by carving out a profitable software business from Evli Group, a Nordic wealth manager. Leveraging Mosaic’s AI-powered sourcing, the team rapidly scanned a universe of over two million companies and pinpointed 200 high-potential targets across Europe. By triangulating Bregal Milestone’s proprietary blend of publicly available data sets with structured and unstructured private data — including 1,500 documents and 350 CRM records related to these specific acquisition targets, the team further refined the list to 25 suitable and actionable priority targets. This data-driven precision enabled the acquisition of two add-ons, Novare Pay and Aktieinvest, within six months of the initial investment.
By allowing the Allshares team to close these deals outside of a traditional process, Mosaic significantly boosted Allshares’ market share, expanded its leadership position in Sweden, and nearly doubled revenue and profitability between 2023 and 2024.[[4]]
"The Bregal Milestone team uses generative AI and their Mosaic platform to turbocharge the value creation plan at Allshares. We harness Mosaic to move faster and go deeper. This allows us to play offense and grow enterprise value on an accelerated basis."
Michael Ingelog
Chairman - Allshares
Developing a data-centric strategy
We see three prerequisites to a successful data and AI strategy in the European midmarket:
Strategic clarity
As a private equity firm, the choices we have made are not without alternatives. A GP can choose a more traditional approach, i.e., gradually increasing the use of data in the business only as much as needed. Equally, an increasing number of off-the-shelf solutions from independent software vendors target the investment industry and, for a while, might confer if not the same value-add, then at least similar marketing benefits. The difference between a white-labelled standard solution and a proprietary platform might not always be evident to the LPs or business owners the GP interacts with.
At Bregal Milestone we are committed to creating a real and lasting investment edge through a proprietary in-house data and AI capability. Our data team is fully embedded in our investment activity, working with both investment professionals and portfolio executives on a daily basis. The use cases Mosaic tackles are selected with the sole purpose of executing and supporting attractive investments and not with a thirdparty software vendor’s view on building a broad but shallow feature set attracting a maximum number of investment firms. Mosaic has been built by Bregal Milestone to constantly scale and improve, and avoids the risks of vendor-lock-in, levelling of the playing field and lack of control that thirdparty solutions would bring.
Our investment in Mosaic will generate a return on investment solely by generating superior investment returns, which we believe best aligns us with our LPs and our portfolio company partners alike.
Data architecture
The backbone of our platform lies in a solid and dependable data structure, allowing each data point to be related to the relevant companies (with source-agnostic integration of company identifiers) as well as industry sectors and sub-sectors (using our proprietary machine-learning based sector classification model). This architecture is enhanced by our best-of-breed search technologies, combining semantic search with machine learning predictions, relational queries and generative AI to provide enhanced search and analysis capabilities.
This multifaceted approach enables our team to query and analyse the entire data ecosystem (public and private) with high precision and without constraints imposed by off-the-shelf data sets. The result has been a significant improvement in the speed and accuracy of our analysis in areas as diverse as underwriting (e.g., company valuation) and value creation (e.g., tuck-in M&A targeting) for our portfolio companies.
Data culture
Data culture is the least understood but most fundamental part of a successful data and AI strategy in private equity. Enforcing data hygiene in an organisation of timeconstrained investment professionals who have seen success operating in a more traditional way will rarely work. Though many firms inside and outside of our industry think primarily of “CRM hygiene” when speaking of data culture, it would be wrong to imply that the path to success lies in enforcing some unpleasant data entry chores with vague long-term benefits.
We had the opportunity to build Bregal Milestone as a data-centric organisation. Each byte of data we ingest as a firm, whether it is a business plan, company presentation, meeting note or industry report, is used to train our Mosaic AI platform, which then powerfully supports and amplifies the work of each sector team or investment professional, perfectly aligning interests across the firm. The team has experienced the benefits of this first-hand and has embraced our data-centric approach as an integral part of their workflows. As a result, Mosaic has a comprehensive, well-structured and, what we believe to be, unique pool of data to create
actionable intelligence from.
Conclusion
Advances in data and AI technology invite private equity managers to take a stance on whether they see possible improvements as incremental or strategically transformational. For a data strategy in private equity to be effective, it needs to be specific to the opportunity set the manager targets, supported consistently with commitment from firm leadership and accompanied by an effective data culture embraced across the organisation. As we invest in an underserved, fragmented and highly dynamic segment of the market, Bregal Milestone will continue to develop and enhance its proprietary AI platform for the benefit of our investors.
