The Right Model for Digital and AI Capabilities: Five Lessons from the PEI Operating Partners Forum 

At this year’s PEI Operating Partners Forum, I was fortunate to be asked to moderate a panel on the right model for digital and AI capabilities. This well attended session focused on the practicalities many portfolio companies are now facing in this area.   

Singulier also co-hosted the London Value Creation Evening alongside Vertice, bringing together a curated group of senior private equity leaders for more candid discussion on what is really driving value creation in private equity today. 

What stood out to me most across these conversations was tha AI is simply not just another chapter in the same digital transformation playbook. Too many leaders still approach it as another technology transformation programme to delegate, phase in and measure later. That instinct is understandable, but in reality, it is off the mark. 

What emerged from the discussion was not, fundamentally, a technology conversation. It was a business strategy conversation. And that distinction matters. 

AI forces first principles, not just process optimisation  

Most technology waves of the past three decades have been about doing the same things faster, cheaper or with greater visibility. AI breaks that pattern. 

It does not simply ask how to improve an existing workflow. It asks a far more difficult question: why does this workflow exist at all? 

That is the real shift. The companies that will create disproportionate value from AI will not be the ones that automate legacy ways of working at the margin. They will be the ones willing to challenge operating assumptions that survived only because changing them once felt too slow, too expensive or too risky. 

In practice, this means AI is not just a productivity layer. It is an opportunity to redesign decisions, roles, handoffs and even where human judgement genuinely adds value. The cognitive centre of gravity starts to move from person to machine. Once that happens, the operating model has to change with it.

The value creation curve is far steeper than in traditional transformation programmes 

One of the clearest themes from the discussion was speed. 

AI is creating a business impact on a different curve from prior digital or technology programmes. This is not a story of steady incremental gains over three years. It is a story of compressed timelines, asymmetrical upside and very real execution risk. 

We are already seeing examples that would have felt implausible not long ago: teams producing a year’s worth of code in a week, service environments collapsing resolution times from months to days, and product or process changes going live at a pace that would once have required an entire annual roadmap. 

For portfolio companies, that changes the board-level question. Leadership teams need to move fast enough to capture the upside before competitors reset the baseline. 

The window is open. But it is narrowing faster than many boards appreciate. 

Technology cannot lead; the business context must 

This is where many digital transformations have gone wrong, and AI will be no more forgiving. 

Too often, the initiative begins in technology, gets framed around tooling, and ends up stuck in a cycle of pilots. The result is activity without economic consequence. 

If AI lives solely in IT, it will usually die in IT. 

The right starting point is not the model, the platform or the architecture. It is the business problem, the user journey and the P&L outcome. That is why AI transformation must be owned by business leaders in partnership with technical teams, not delegated downward as a specialist programme. 

Functional leaders, operators and technologists need to be in the same room early, aligning on what value looks like before anyone starts building.  

L-R: Andrew Dawson , Marine Diot (EQT), Gilad Amir (Pollen Street Capital), and James Craig (GCP)

Ruthless prioritisation is non-negotiable 

AI creates possibilities faster than most organisations can absorb them. That is precisely why focus matters. 

Capabilities are evolving every few months. New tools appear constantly. Vendor noise is relentless. In that environment, broad ambition is not a strength. It is a liability. 

The consensus is that “doing AI” everywhere at once poses a real risk to sustainable transformation. The best approach is to identify the small number of use cases where AI can deliver a structural, measurable impact and then going deep. 

That usually means prioritising the places where one of three conditions is true: 

Everything else can wait. 

Without that discipline, companies build a graveyard of pilots: technically interesting, commercially marginal and quietly abandoned. Focus is not a constraint on AI strategy. In most cases, it is the strategy. 

Let’s Talk