Two portcos run the same AI playbook. One compounds. One unravels.
In the board meetings I sit in, one question keeps landing on PE operating partners from their fund’s leadership: is our portfolio AI-ready?
It’s the wrong question.
The right one: can the system underneath each portfolio company hold what AI amplifies?
The Amplification Problem
AI is the most significant single amplifier most portfolio companies have introduced into their go-to-market systems in the past two years. It accelerates pattern recognition, decision support, and execution at scale. What AI doesn’t do is build the system underneath.
When the underlying business reads itself accurately, AI compounds advantage. Forecasts tighten because the read on the business is shared across functions rather than disputed. Pattern recognition speeds up because the company has already decided which patterns it’s looking for. Decisions are calibrated to enterprise economics in something close to real time.
When the underlying business doesn’t read itself accurately, AI accelerates the same things in the wrong direction. Forecast distortion compounds faster. Volume-over-quality customer acquisition scales faster. Capital misallocation hides behind cleaner-looking dashboards. The portfolio company runs harder, not smarter.
The same AI deployment, in the same vertical, with the same vendor, produces opposite outcomes depending on what sits underneath it.
What That Actually Looks Like Inside a Portco
It rarely shows up as anything dramatic. The signs are quieter and more familiar than that.
Marketing, Sales, and CS each report to the operating partner with their own working definition of an “engaged” account, and the operating partner makes investment decisions based on three definitions stitched together into a single number. Pipeline coverage looks consistent quarter over quarter, but the rules for stage progression have drifted within the team to keep the optics steady. The customer cohorts in the renewal model look strong on the surface, but the depth of product usage beneath them has been trending the wrong way for two quarters, and nobody flagged it because the dashboard wasn’t built to catch that distinction.
In each case, the data is there. The functions are reporting in good faith. The numbers are technically accurate. The business underneath isn’t built to reconcile what those numbers mean to each other.
AI doesn’t fix this. AI runs faster on top of it. The forecast gets more sophisticated. The predictions get more confident. The actual decisions still rest on a stack of definitions that don’t line up.
What This Means at Diligence
The risk that has emerged within the extended hold period is what we call valuation double jeopardy: reduced EBITDA compounded by a reduced multiple during diligence. Both inputs move against the asset at the same time, each reinforcing the other.
AI deployment sharpens both directions.
In a portfolio company where the underlying business is reading itself accurately, AI becomes a defensible operating advantage that holds up under diligence. The buyer’s team can pressure-test the forecast, and it survives. The story is compounding, not promising.
In a portfolio company where it isn’t, AI investment surfaces in three specific ways during diligence.
- Forecast distortion the buyer’s diligence team can quantify against actuals.
- Capital misallocation visible in CAC trends that don’t improve despite AI tooling.
- Revenue quality questions the buyer can pressure-test against cohort behavior.
These patterns aren’t subtle to a sophisticated buyer. They are exactly what diligence is looking for. The reduced multiple at exit is the cost of running AI on top of a business that wasn’t reading itself accurately to begin with.
The Strategic Question
The strategic question isn’t whether to deploy AI inside the portfolio. That capital is committed. The vendor selections are made. The deployment timelines are set.
The strategic question is whether the system underneath each portco can hold what portfolio AI deployment is about to amplify, and whether that’s true for every asset in the fund or only some of them.
That question is answerable, with the company’s own data, before the next AI investment cycle locks in.
If you’re approving 2026 AI investment plans across the portfolio and you don’t yet have a clear read on which portcos are positioned to compound from the deployment and which are positioned to combust, that’s worth answering before the capital deploys, not after the buyer’s diligence team answers it for you.
The architectural concepts and methodology behind this argument are the proprietary work of Marketing Affects.