Industry insights
LLMs and the AI-powered portfolio: 5 takeaways from the IPEM Value Creation Summit
LLMs and the AI-powered portfolio: 5 takeaways from the IPEM Value Creation Summit
OMMAX and Singulier were pleased to sponsor the Value Creation Summit at IPEM Global 2026, one of Europe’s largest private markets gatherings.
Toni Stork, Group CEO of OMMAX and Singulier, hosted the panel LLMs and the AI-powered portfolio, joined by Emmanuel Cassimatis of SAP, Gaël Gibert, Senior Director, Digital & AI at Advent International, and Adam Nahari, Operating Partner at Apollo.
The discussion focused on the key questions GPs are asking as LLMs move into the portfolio: where can they create measurable value, how should AI be assessed in diligence and post-close plans, and what governance, talent and operating model are needed to move beyond pilots?
Here are five key takeaways from the panel.
1. AI assessment should start in diligence
AI risk and opportunity should be assessed as part of every investment process, with the depth of analysis adapted to the target’s exposure.
A robust assessment should consider the sector’s AI potential, the company’s capabilities, the gap between current maturity and potential, and the risk of competitive disruption. The analysis should also distinguish between AI capabilities that will become table stakes and those that could create durable differentiation over a five- to ten-year hold.
The quality of a target’s proprietary data, models and infrastructure should also be tested. These assets may determine whether AI creates a defensible advantage or simply brings a company up to market standard.
2. Governance and organisational readiness are critical
Successful AI transformation depends less on the technology than on leadership alignment, ownership and the ability to manage change.
Post-close can be the right moment to deepen the assessment, when management access is stronger and the transformation can be scoped properly. A clearly defined investment envelope and delivery timetable are essential. Without them, an apparently contained initiative can expand significantly as infrastructure, data and system dependencies emerge.
The transformation also needs broad executive alignment. AI modernisation affects systems, processes and operating models across the business, making it difficult to succeed when ownership sits with only one function.
3. Focus on a small number of commercially important use cases
Broad experimentation rarely creates meaningful impact. The stronger approach is to prioritise one or two use cases directly linked to how the business creates value.
Potential areas include customer-facing propositions, product innovation, inventory optimisation and operational efficiency. The starting point should be the desired customer experience or business outcome, not the capabilities of a particular model.
There is also an important strategic distinction between doing things better and doing better things. Incremental efficiency improvements can deliver value, but new products, customer experiences and revenue models have the potential to create a more significant shift in the economics of the business.
4. Design for flexibility rather than vendor dependency
The LLM landscape is evolving too quickly for companies to rely too heavily on a single model provider.
Vendor lock-in can create pricing exposure, limit flexibility as capabilities change and prevent companies from using the best model for each task. Architectural choices and data ownership are therefore as important as partnership decisions.
This also applies to solution design. Using proprietary memory, storage or workflow features can create dependencies that are difficult to unwind later. Flexibility needs to be built into the architecture from the beginning.
5. Prove the economics before scaling
AI transformation should be measured through demonstrated business outcomes rather than projected productivity gains.
The recommended approach is to prove return on investment through an iconic use case before expanding across the organisation. This avoids premature scaling and creates a stronger basis for securing further investment.
The strongest immediate opportunities appear to be areas such as agentic software development and customer-service automation, where productivity improvements are sufficiently evidenced to inform decisions. Other workforce and efficiency cases require a more cautious, company-by-company assessment, with action based on demonstrated rather than projected returns.
The central message: LLMs are not the value-creation strategy. They are the enabler. The value comes from selecting the right business problem, aligning the organisation behind it and proving measurable impact before scaling.
Related content
