What AI roll-ups look for in acquisitions, and what that means for your clients

The plain-language answer

An AI roll-up buys established services businesses and rebuilds how the work gets done, using AI software for tasks people used to do by hand. Its acquisition criteria follow directly from that thesis: firms where delivery is labor-heavy, repeatable and documented, where clients stay for years, and where the margin gained from automation is large enough to pay for the purchase.

Two consequences matter for an M&A advisor. Clients in those sectors may receive approaches, and the thing that makes them attractive, years of documented workflows, is the same thing AI labs and data buyers license. Those are separate transactions. A license does not require a sale.

The six criteria AI roll-ups screen for

Acquirers rarely publish their criteria, but the logic of the model points to six screens. Use the last column to test a client before an approach arrives.

The criteria also explain why roll-ups insist on keeping both the clients and the people who know the workflow through the transition. An acquisition that delivers one without the other does not support the thesis.

Which sectors fit the profile

This page does not track who has bought whom; deal announcements move quickly, so rely on dated reports from the acquirers or reputable press for that. What can be said is which sectors match the criteria, and what licensing would and would not involve in each.

Why the same records interest AI data buyers

AI development is shifting from models that answer questions to agents that carry out multi-step tasks. Training and evaluating those agents takes records of real work: requests, decisions, tool use, exceptions and outcomes. That material lives inside companies and is thin on the public web.

Researchers at Epoch AI estimate that if current trends continue, language models will fully use the stock of public human-generated text sometime between 2026 and 2032, a forecast with wide uncertainty. Permissioned, rights-cleared business records are one answer to that constraint.

The roll-up and the data buyer want different things from the same firm. The roll-up wants to own and run the workflow. The data buyer wants a licensed copy of the records that show how the workflow ran, under agreed de-identification and redaction rules, while the firm keeps ownership.

The either, both or neither rule

When a client in a roll-up sector asks what its options are, sort it into one of four rows.

The third row needs the most care. A license signed before a sale should be disclosed to bidders and fit the deal; a sale signed first hands the licensing decision to the buyer. Our comparison on whether to sell a firm to an AI roll-up walks an owner through that choice, and AI roll-up vs traditional PE roll-up explains why AI buyers tend to retire legacy systems faster.

What this means for an advisor making the introduction

The advisor's part is the introduction; the company does the rest with SourceX.

1. Ask the owner whether they want to hear about licensing, and say plainly that you may earn a referral reward.
2. Share your referral link so the company applies at sourcex.si/apply with your code attached, or submit it through the referral form.
3. SourceX checks size, operating history, breadth of records and rights with the company's sponsor.
4. The company builds a data inventory: each system, how many years it covers and what can be exported.
5. Price and terms are agreed with the company before any buyer sees the opportunity.
6. AI labs and data buyers review it; once a company is deal-ready, buyers typically respond within about two weeks.
7. When a deal closes, the data is delivered under the agreed rules and the company receives a one-time payment, typically within about 60 days of invoicing once the buyer selects the data.

You never export, upload or describe confidential records. Partners earn 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, capped at $100,000 per referred company; rewards become payable only after the buyer pays and SourceX receives its fee, and no reward is guaranteed. Advisors with licensing duties should check their own rules on referral fees and disclosure. The page for M&A advisors covers sell-side mandates, and buy-side firms have their own playbook for targets that are not for sale.

Limits and open questions

• Supply is large, but not every firm is a target. McKinsey estimated in February 2026 that about six million US small and medium-size businesses will face ownership transitions by 2035, of which more than one million are viable candidates for sale. Most will never meet a roll-up's criteria or a licensing baseline.
• Automation margins are a forecast. Whether AI delivery produces the margins an acquirer underwrote is not yet proven across sectors, which is why earnouts and rollover equity deserve scrutiny.
• Legacy records can disappear. A roll-up that migrates an acquired firm onto its own platform may retire the systems that hold years of history unless someone preserves exports first.
• Licensing has its own gates. Nothing is binding until the company agrees price and terms and signs, and a company can fail the rights review even when its records are deep.

Next step

List the clients in roll-up sectors that also clear the licensing baseline on the who qualifies page; the network opportunity finder helps you work through the list. When one fits and the owner agrees, register as a partner and make the introduction.