AI readiness assessment for portfolio companies: start with owner dependency

An AI readiness assessment for a founder-led portfolio company starts with one question: does the business run without the founder? If decisions, processes and data still live with one person, AI has nothing stable to work on. We assess that first, then say where AI is ready now, ready after fixes, or not advised.

We install systems in companies that make them run without owners and ready for automation.

What standard diligence misses in a founder-led company

Financial and commercial diligence tell you what the company earned. They don't tell you how it runs. In a $10M to $100M founder-led company, three things usually show up after close:

• Decision authority. Pricing, hiring, refunds and vendor calls still go to the founder, because nobody else was ever allowed to make them.
• Unwritten process. The way work gets done lives in a few people's heads. When one of them leaves, the process leaves too.
• Untested data. You open the hood and get an Excel file and a QuickBooks login. Nobody has checked whether the numbers can carry a dashboard, let alone an AI agent.

None of this is in the data room. All of it hits the first 100 days.

Owner dependency is a valuation risk

The question I ask in every engagement is simple: if the owner took two months off, what happens?

Often the answer is one word: chaos. That answer is the risk. A buyer pays for earnings the business can repeat. When the earnings depend on one person being in the room, the buyer is paying for that person, and that person usually plans to leave.

You can measure it. List the decisions the owner still makes. Note which processes nobody else can run. Check which numbers only the owner trusts. The longer those lists, the bigger the discount.

The operating diligence we run

Every assessment runs on SCOPE™ by Efficiency Sprint, so every portfolio company is measured the same way and the results compare across the portfolio.

The interviews are short: two hours with the owner and 15 minutes with each key person.

Decision authority

A map of what the owner still decides, and what each manager is actually allowed to decide. This becomes the first draft of a decision authority matrix: for each decision, who may make it and up to which limit.

Process and SOPs

What's written down, what isn't, and what's written and not followed. Our library of 150+ pre-made SOPs is the starting set, so the team edits instead of starting from a blank page.

Data tests

Management can give you beautiful reports, show you beautiful numbers, even tell you everything is great, and your gut is telling you: no, it's not. The root cause is always how correct and trustworthy the data is.

We go to the sources, not the reports built on them. How does the data get recorded? Is it updated? Is every department tracking? Can it carry a weekly dashboard, and can it carry an AI agent? I make everyone report. Without documented proof of what people do at work, I don't know how to manage them, and neither will the new owner.

AI and automation potential

For each area, one of three answers:

• Ready now: the process is written, the data is clean, someone owns the outcome.
• After fixes: the work is worth automating once the process or data is repaired.
• Not advised: the cost of a mistake is too high, or the volume too low.

That's the chapter most AI due diligence starts with. We put it last, because it depends on everything above it.

From findings to a management system

The assessment ends with what to do in three to six months, not in three to five years. In the first 100 days after close, that usually means:

1. A decision authority matrix, so decisions stop waiting for the founder.
2. A weekly operating rhythm, with the numbers on one page every week.
3. Written SOPs for the processes the business can't afford to lose.
4. Dashboards on tested data, and only then the first AI agents, each with one owner and a limit.

What it looked like in practice

In a construction company I work with, everything ran through the owner and nobody knew their role. We built the system department by department:

• The sales pipeline went from $1M to $14.5M in three months.
• A-players were hired in about a month.
• $1.2M of $1.4M in overdue receivables was collected once one person owned receivables.

That's what was done in one company, not a promise for the next one. The order is the part that repeats: right people in the right places, goals that match the company's goals, then measure everything.

Change management is painful. From the inside, it can look like chaos for a while. The key is to stay consistent. My playbook: grow the team for a while so the old team's knowledge moves to new people without a threat, give two to four months of overlap if the budget allows, then decide by data who moves to the next stage of the company.

Next step

See how this works for a fund: Efficiency Sprint for private equity. Or start with one company: the Diagnostic.