Thinking on AI, Private Equity, and Defensible Value

The first 100 days of a hold period are the only time a portco's leadership team is genuinely open to structural change. After that, the organization finds its rhythm, and change gets harder. If you're going to sequence AI value-creation work, this is the window.

Most operating partners know this. The problem isn't intent — it's sequencing. AI initiatives launched in the wrong order create noise before they create value, and noise in the first quarter of a hold period is expensive.

Here's the sequence that works.

Days 1–30: Diagnostic. Before you recommend a single AI initiative, you need a clear-eyed view of the portco's current state — data quality, infrastructure maturity, talent gaps, and the specific business problems worth solving. This is not a technology audit. It's a business audit that happens to include technology.

Days 31–60: Prioritization. Map every AI initiative to a specific EBITDA lever — cost per unit, revenue per customer, cycle time, margin. Rank by impact and implementation risk. The output is a prioritized initiative list with a financial model attached to each item. If you can't attach a financial model, the initiative doesn't make the list.

Days 61–100: First initiative in production. One initiative, fully implemented, with a measurement framework in place. Not a pilot. Not a proof of concept. A production capability with a scorecard that the board can read.

The one-page playbook template that maps this sequence — with the financial model structure, the initiative scoring matrix, and the board reporting format — is available through the 100-Day AI Value-Creation Playbook engagement. It's the same document we deliver to every operating partner we work with.

If you're entering a new hold period and want to run this sequence with a structured framework behind it, that's what the 100-Day Playbook is built for.

Exit diligence has changed. Two years ago, a buy-side team might ask one or two questions about AI. Today, the AI section of a quality of earnings engagement is standard. The questions are getting more specific. The discounts for bad answers are getting larger.

Here are the six questions every buy-side diligence team is now asking — and what a defensible answer looks like for each.

1. What AI capabilities are in production? Not "what AI do you use" — what is in production, generating output that affects business decisions. The answer should include a model inventory with materiality classifications.

2. Who owns the data those models run on? Third-party data dependencies are vendor risk. Proprietary data is a moat. Buyers price the difference.

3. What is the governance structure? Who owns each model. Who signs the risk acceptance. What the incident response process looks like. A company that can answer this question has a managed AI capability. A company that can't has a shadow capability — and shadow capabilities are discounted.

4. What does the monitoring framework look like? Pre-deployment testing, ongoing performance monitoring, drift detection. The controls that separate an AI capability from an AI liability.

5. What is the regulatory exposure? State privacy laws, sector rules, the federal patchwork. A company that has mapped its AI capabilities to its regulatory exposure is defensible. A company that hasn't is a remediation cost estimate waiting to happen.

6. What does the board reporting look like? The one-page monthly view. The quarterly narrative. The language a regulator will accept. If the board hasn't been reading AI reports, the capability hasn't been governed.

The Exit-Readiness AI Attestation is a four-week engagement that answers all six questions, produces the documentation a buy-side team will ask for, and delivers a board-ready attestation you can put in the data room. If you're 12–18 months from a process, this is the engagement to run now.

LP letters have always been a proxy for how a firm thinks. The sections that get written carefully signal what the firm believes matters. The sections that get skipped — or filled with boilerplate — signal the opposite.

Most firms haven't written a real AI section yet. A few have added a paragraph about "exploring AI opportunities across the portfolio." That paragraph is now a signal — just not the one they intended.

Here's what a real AI section in an LP letter looks like.

It starts with a framework statement. Not "we are exploring AI" — a statement of how the firm thinks about AI as a value-creation lever. What the firm's position is. What the firm's process is. What the firm's standard is for AI capabilities in portfolio companies.

It continues with portfolio-level reporting. Which portcos have AI capabilities in production. What those capabilities are mapped to in the value-creation plan. What the governance structure looks like. This is the section that separates firms that have a portfolio-wide AI program from firms that have a collection of portco experiments.

It ends with a forward-looking statement. What the firm is building toward. What the standard will look like at exit. What LPs should expect to see in the next letter.

The firms that write this section well are raising on it. The firms that skip it are leaving a question in the room that sophisticated LPs are starting to ask out loud.

The Portfolio-Wide AI Readiness Program is the firm-level engagement that gives you the material to write this section — and the portfolio-level data to back it up. If your next fundraise is in the next 24 months, this is the program to run now.