Phil discusses AI’s impact on fintech style private markets, from deal sourcing to data architecture, and why unified data drives alpha.
Phil, could you start by walking us through your professional journey and how it has shaped your perspective on AI in private markets?
I spent 15 years at BAE Systems and IBM, including 2012–2016 on the IBM Watson team, helping usher in enterprise AI. In those early years, it was about finding the use cases and business applications for Machine Learning and Natural Language Processing. One thing that was always very apparent in the age of AI, was the need for organizations to build up their proprietary data to enable the AI to drive competitive advantage. I also saw how Data Engineering represented 90% of the challenge of operationalizing AI.
I cofounded Deal Engine (originally called Filament) in 2016 and have spent the last decade applying the ‘full stack’ engineering of Market Data technology for the Private Equity (PE) industry, codifying the strategies and rolling with the ever advancing underlying technologies. We’ve built up a 50-strong team of technical and market data experts.
From your vantage point working with PE firms across both sides of the Atlantic, how would you describe the current state of AI adoption in deal sourcing and market coverage?
There is a spectrum of sophistication and adoption across the whole PE market. Outside of a 10% of advanced PE firms, the vast majority are figuring out how to operationalize their AI program. This has been further challenged by the rapid advancements of Large Language Model (LLM) technology, and the now compelling need to have a strategy to remain competitive. Furthermore, it is becoming a key topic during Limited Partner (LP) fundraising. In the last six months, there has also been an understanding shift in the market, led by the example of the leading AI-powered PE firms, that their long-term AI program must be built on strong data architecture foundations. Essentially, a firm must integrate their proprietary knowledge to allow the rapidly evolving LLM tools to flourish. The next 30% of firms are now putting the architecture in place. The rest are still figuring it out.
What are the most common data engineering and architecture issues you see private equity firms struggling with today?
At an architectural level, the firm needs to integrate all the different systems internal and external that provide different facets of the firm’s market knowledge and intelligence. As conventional API interfaces are being blended with the newer Model Context Protocol connectors, the challenge remains of getting systems to coalesce into a single, curated system of record. The data is a combination of unstructured and structured and competing datapoints must be interpreted, linked and resolved. One particular challenge that has plagued the industry has been Entity Matching to align all the relevant data for an individual company across multiple inputs. Furthermore, the combination of structured and unstructured data requires careful decision-making across different database technologies, blending performance and the cost of ownership.
Fundamentally, most PE funds struggle by not having a Data Engineering function to resolve these complexities and solve practical problems along the way. Similar challenges lie in the complexities in DevOps, Quality Assurance and Database Engineering. Ninety nine percent of firms don’t have all these bases covered and so need help to build their engine.
How does fragmented data across CRMs, VDRs, emails, PDFs, and internal systems limit the effectiveness of AI in practice?
AI pointed at one fragmented system may provide some tactical advantage. We’ve seen all the individual systems offer AI tooling. But to truly gain alpha, the firm must coalesce all the knowledge of a firm into a centralized data layer. Pointing AI at this enables full intelligence and insight on a target in one 360-degree view.
Which insights from the Citi Institute’s view that AI is only as strong as its underlying data strategy resonate most with what you’re seeing on the ground?
I couldn’t agree more with the article. The firm leadership needs to have a strategy for their data architecture. It doesn’t necessarily need to be tech leadership in the firm, although this helps, but fundamentally this can be the vision of the managing partner. The strategy of the leadership must be to to coalesce their firms’ proprietary knowledge and edge into one layer, and then roll out a program of LLM-based work to derive increasingly more powerful operational edge. This program can then be executed with a blend of internal and external components and partnerships.
Given your experience at IBM Watson, what are the biggest misconceptions PE firms have about deploying AI versus the reality of implementation?
PE professionals understand business models, so the concept of a unified knowledge layer and its path to alpha is widely understood. As a professional body, they will naturally stay live in the market, so are well-informed on the forefront of Frontier AI technology potential.
However, the collective blind spot in most PE firms lies in understanding the challenges are in the data and backend engineering, as discussed, this being the key foundational capability that allows the rapid evolution and adoption of the latest Frontier AI.
Where are you seeing AI and data platforms genuinely create competitive advantage in private equity today? Are there specific use cases that stand out?
I’m seeing AI have dramatic impact on the efficiency and efficacy of sourcing and market monitoring, allowing firms to cover more of the market, spend their time on the best fit deals for their strategy and go into meetings super-prepared with the power of five (virtual) analysts supporting each dealmaker in meeting prioritization and prep. Finding and closing the best deals is the life blood of a General Partner (GP). Baking the value-creation strategy into that sourcing leads to higher returns on the output. This is putting AI at the very core of a GP’s raison d’etre, not just efficiency gains but step-change efficacy gains.
Why do so many deal-tech initiatives stall or fail to move beyond spreadsheets and disconnected point solutions?
Strategy, accountability and choosing the right external partners and providers are essential foundations. With those three in place, you have a real chance of success. The remaining, crucial component is leadership that can drive change and actively incentivize the right behaviors. The world is changing, and the leading PE firms will be those whose leaders are proactively driving that change within their organizations.
For firms looking to move beyond fragmented tools, what does a strong, scalable data and AI architecture actually look like in practice?
Build a data layer that integrates your internal proprietary knowledge systems (CRM, Sharepoint, Email, VDR) with a healthy mix of external market data, blending structured data (financials, Net Promoter Scores (NPS), firmographics) with unstructured context (news, market opinions, leadership track record, founder intent). This layer becomes the integrated knowledge and memory layer of your firm, enabling the 24/7 agents working for your firm. It also provides the fuel for your Claude or ChatGPT interface, enabling your team to query and interact with your unified knowledge base.
What practical advice would you give to private equity leaders aiming to successfully harness AI for deal sourcing and market intelligence over the next few years?
Build a strategy and architecture that is future-proofed for the inevitable rapid development of Frontier AI. Think of elements of your architecture that persist, such as unified knowledge and the intelligence-dealmaker interaction. Be flexible to switch out LLM providers as they jostle for supremacy. Find vendors that represent the best piece of your architectural jigsaw puzzle. And work with technical partners that can help you on that journey and ride the wave of AI that will transform everything in the industry.
Stay Ahead of the Financial Curve with Our Latest Fintech News Updates!
