How PE Operating Teams Should Assess AI Readiness in a Portfolio Company

You do not need a maturity model with forty criteria. You need five questions that let you segment a portfolio company in one conversation and decide whether a 90-day sprint will produce a result worth funding. Here they are.

Every operating partner is being asked the same thing right now: what is AI doing for the portfolio? The honest answer for most firms is that it varies wildly by company, and there is no consistent way to compare them. One company shipped real automation. Another ran a pilot that stalled. A third has not started. Without a common framework, you are comparing anecdotes.

The goal of an assessment is not a score for its own sake. It is to answer one question per company: is there a workflow here where a focused, fixed-price sprint would move an operating metric inside the hold period? These five questions get you there fast.

Question 1: Is there one workflow with visible, quantifiable cost?

Not a category. One workflow. The monthly close, invoice processing, intake routing, reconciliation, customer-operations follow-up. The test is whether the company can put a number on it today: hours per week, days per cycle, or FTEs tied up.

If the answer is a specific number, you have a baseline and a target. If the answer is vague, the first work is measurement, not AI. Companies that cannot quantify the cost of a workflow are not ready to automate it.

Question 2: Is there an owner who will run the new process?

AI initiatives fail on people far more often than on technology. The question is whether there is a named operator, usually a COO, controller, or function head, who owns the workflow and will own the new version of it.

A company with an engaged owner is ready. A company where "IT will handle it" or no one clearly owns the process is a red flag, regardless of how good the data looks. Adoption needs an owner with skin in the game.

Question 3: Is the data available and good enough?

Not perfect. Good enough. Is the input the workflow depends on accessible, structured enough to use, and not locked inside one person's head or a system no one can reach?

You are not looking for a clean data warehouse. You are looking for the absence of a blocker: data that is fundamentally unavailable, or so chaotic that the first six months would be a data project. If the input is reachable and consistent, the company clears this bar.

Question 4: Can a new system reach the stack the team already uses?

The systems that get adopted live where people already work: the ERP, the CRM, the ticketing tool, the spreadsheets. The question is whether a new AI system can integrate with that stack, or whether the company is so locked down or so fragmented that integration alone would consume the engagement.

Most mid-market stacks are workable. The ones to flag are heavily customized legacy systems with no integration surface, which turn a 90-day sprint into a multi-quarter platform fight.

Question 5: Is leadership willing to measure a before-and-after?

The whole point is provable impact. Is the CEO willing to set a baseline at kickoff and be held to a measured result at close? Or do they want a pilot they can point to without committing to a number?

A leadership team that will commit to a measured outcome is ready to create value. One that wants activity without accountability will produce spend, not leverage. This question often matters more than the four technical ones combined.

How to use the answers

Score each company yes, partial, or no across the five questions. The pattern tells you where to act.

Ready now (mostly yes): a quantified workflow, a real owner, usable data, a reachable stack, and a CEO who will commit to a number. Fund a 90-day sprint here first. This is where impact lands inside the hold period.

Fixable (yes on people and leadership, gaps on data or systems): the will is there but a blocker exists. A short readiness step clears the path before the sprint. Sequence these second.

Not yet (no on owner or leadership): the gap is organizational, not technical. No model will fix it, and a sprint will stall. Address ownership and accountability before spending on AI.

Run this across the portfolio and you get what most operating teams are missing: a consistent, defensible read on where AI will create value, and a prioritized order to go after it. That is exactly what the Portfolio AI Readiness Diagnostic produces at scale, and what the 90-Day AI Operations Sprint delivers against once the priorities are clear.