In the first 100 days after a deal, AI helps in three narrow places and hurts in one. It helps you find out who already uses AI tools and for what, it helps people read and reconcile documents and reports so they decide faster, and it works best under one simple rule about data. It hurts when you buy an AI platform before your monthly reporting is stable. This article is for the CEO, CFO, private equity operating partner or integration lead who wants to know where AI earns its place in the months after a deal, and where it is a distraction.
I was COO and then President of a public company that closed more than 40 acquisitions, and I ran the integration on most of them. Integration is mostly people keeping promises, and the useful question about AI is which of those promises it makes easier to keep.
What integration work actually consists of
For the first few months, integration comes down to five kinds of work:
- People. Who stays, who leads what, and what you tell everyone and when.
- Promises. What the new owner, the lender and the customers were told would happen.
- Reporting. One set of monthly numbers that the owner, the lender and the managers all trust.
- Contracts. Consent requirements, notice dates and terms that the deal may have triggered.
- Systems. Which software, vendors and processes stay, change or go.
Most of this is judgment: deciding who leads a team, what to say to a worried customer, which promise to renegotiate. AI does not do that work and should not be asked to.
A smaller part is reading. Someone has to read two sets of contracts, compare two charts of accounts, reconcile two versions of the same report, and summarize what changed since last month. That is where AI can give you hours back. The full 100-day sequence is laid out in the first 100 days after the deal, and the one-page version is the first 100 days checklist. The rest of this article covers only the AI part.
The three places AI fits
1. Finding out who already uses AI tools, and for what
Picture a mid-sized company that was just acquired (a made-up composite, not a client). Nobody in leadership has approved any AI tool. Yet a customer service lead has been pasting complaint emails into a public chatbot to draft replies, and a finance analyst has been using one to tidy a spreadsheet of customer prices.
Neither is hiding anything. Both are probably among the more capable people on the team. But the new owner now has customer and pricing information sitting in tools nobody approved, and nobody has asked the question.
The first job is simply to ask, in a friendly way, who uses these tools and for what. Ask, do not audit. People who feel audited stop talking. The practical way to find those people is in the guide to spotting your AI champion. Owner: the COO.
2. Reading and reconciling documents and reports
Picture a CFO a few weeks in who has two monthly reports from two companies, built from two different definitions of revenue, plus a stack of customer contracts from the acquired business. Someone has to line up the figures, flag where definitions differ, and pull out which contracts mention a change of ownership.
An AI tool can do a first read of all of that in an afternoon and produce a short list: these figures do not match, these contracts mention consent. A person then checks the list and decides what it means. The tool saves the reading, and the judgment stays with the CFO.
The rule that makes this safe is simple: AI produces the draft, a named person signs the final. Owner: the CFO for the reports, the General Counsel for the contracts.
3. One data rule
Picture a sales manager who, soon after closing, pastes the acquired company's price list into a free online tool to speed up a quote. It takes seconds, and the price list is now outside the company.
One sentence prevents this: no customer or pricing data goes into any AI tool unless the company has approved that tool. Put it in writing and send it to everyone, including the people who joined through the deal, who have never seen your rules. Owner: the General Counsel, or the CEO if there is no General Counsel.
The one place AI does not fit yet
Do not buy an AI platform in the first 100 days, and in particular not before monthly reporting is stable.
Picture a company that signs a contract for a platform that promises to forecast sales and flag risks. The platform needs clean history to learn from. But the two merged companies still define "customer" and "revenue" differently, and the monthly report changes shape every month. The platform produces forecasts with confidence, and no one can tell whether they are right. People start to doubt the numbers and the tool at once.
The sequence matters. First one trusted monthly report, then clear owners for each number, then a conversation about what a tool should do with them. Buying in the other order gets you wrong answers faster, and it takes leadership attention away from the promises that matter most in those first months. Owner: the CEO, who is the one person who can say no to the vendor.
What the diligence team will ask about AI
A private equity firm or its operating team now asks about AI use, both before a deal and in the months after. Expect five questions. The plain answers an owner can give are below.
- Where is AI being used in the company today? Answer with what you have learned from asking. "We asked everyone. Here are the places it is used, and who is accountable for each." An honest list is better than a claim of "nowhere."
- Who is accountable for it? Answer with one named person, not a committee. Usually that is the COO or CEO, with the General Counsel owning the data rule.
- What data goes into these tools? Answer with the rule: no customer or pricing data unless the tool is approved. Say how people were told.
- What will you never let AI do on its own? Answer with a short list. A good start: nothing that sends money, signs a contract or makes a promise to a customer without a named person's approval.
- What would you do if it went wrong? Answer with who finds out, who decides, and how quickly the owner would be told.
If the board or sponsor wants these answers on paper, they can be written down in a page or two before anyone asks. The agenda and the ten questions are on one page: AI board briefing template.
Who owns AI in the integration
There is no single right role, but each role has a natural piece:
- CEO. Decides what is bought and what is not, and holds the line on the platform question.
- COO. Runs the "who is using it and for what" conversation, and owns the answers to the board.
- CFO. Owns the reading and reconciling of reports, and the sign-off on the numbers that come out.
- General Counsel. Owns the data rule and the contract review, and the list of what must never be built.
- The one person people already ask. In most companies there is someone colleagues go to when a spreadsheet will not behave. Find that person early. They are the best guide to what is already happening, and they can show others how to use a tool well.
The risk is giving AI to the head of IT by default. Deciding who is accountable and what is off limits is a management decision, and it belongs with the people who own the promises.
Mistakes that repeat
- Auditing instead of asking. People who feel watched stop sharing, and the quiet use continues.
- Letting the tool decide. A draft from AI is a draft. Someone with a name signs it.
- Giving the data rule to nobody. A rule with no owner is not a rule.
Frequently asked questions
Can AI speed up post-merger integration?
Yes, in the reading. It can do a first pass over contracts, reports and meeting notes in hours instead of days, so your people spend their time deciding. It does not speed up the decisions themselves, such as who leads what or what to tell a customer.
What should a PE firm ask about AI in due diligence?
Five things: where AI is used today, who is accountable, what data goes into it, what it is never allowed to do on its own, and what the company would do if something went wrong. Plain answers to those five tell you more than a list of tools.
Should we buy an AI platform right after an acquisition?
Not yet. Wait until your monthly reporting is stable and the merged company agrees on its definitions. A platform built on numbers nobody trusts produces answers nobody trusts, only faster.
Who should own AI during integration?
One named person is accountable, usually the COO or the CEO. The CFO owns the reports, the General Counsel owns the data rule, and the CEO owns purchases.
What data rule should we set for AI tools?
One sentence: no customer or pricing data goes into any AI tool unless the company has approved that tool. Write it down, send it to everyone, and name who owns it.
Put the answer on paper
If your board, lender or new investor has asked for your AI position, the AI Board Briefing is a fixed-fee written plan that puts the answer on paper: where AI is already in use, who is accountable, and what should never be built.