Why a single, adaptable operating model beats letting every portco run its own pilot.
Key takeaways
It's common for a portfolio to end up with ten companies running ten different AI pilots, each chosen and managed independently, with no shared metrics and no way to compare results. The fund ends up with a lot of activity and very little that's actually reusable.
This scattered pattern usually stems from good intentions, letting each management team move fast, but it trades away the fund's biggest structural advantage: the ability to apply what works at one company across the whole portfolio.
A shared playbook doesn't mean identical tools everywhere. It means a consistent set of principles, how initiatives get prioritized, how ownership is assigned, what gets measured, applied by each company to its own systems and starting point.
This distinction is what lets a fund compare results meaningfully across a portfolio with very different levels of technical maturity.
The specific tool a manufacturing portco uses for quoting won't be the same tool a services portco uses for intake, and that's fine. What should stay consistent is how each company decides what to prioritize and how it reports progress.
This balance, shared structure with local adaptation, is what separates a genuine portfolio strategy from a loose collection of independent experiments.
Without a central owner at the fund level, even a good playbook drifts as each portfolio company quietly adapts it into something unrecognizable. That owner doesn't need to run every rollout, but they need visibility into how each company is applying the shared model.
This central ownership is also what makes it possible to answer an LP's portfolio-wide question with an actual answer, rather than ten separate stories.
A 30-minute call is enough to tell you whether AI pays for itself here.
