Due diligence has always been a race between the volume of information and the time available to understand it. AI changes that balance. It does not change what diligence is for.
What AI changes. The most immediate shift is in reading. Data rooms contain contracts, financials, policies, correspondence, and operating documents that once took teams weeks to review. AI tools can now summarize, compare, and flag items across large document sets far faster. Unusual contract terms, inconsistencies between documents, and missing items can surface earlier.
That speed matters. It gives the deal team more time for the questions that actually determine value, instead of spending that time finding them.
AI also makes it practical to ask a different kind of question: not only what the documents say, but how the business actually operates. Where does work depend on manual effort? Which processes are held together by specific people? Where are the systems disconnected? Those operating realities often shape post-close performance more than any single contract clause.
What AI does not change. It does not decide what matters. A summary is only as useful as the person who knows which findings affect the investment thesis and which are noise. That judgment still comes from people who understand the business, the market, and the deal.
It does not verify truth. AI works with the information it is given. If documents are incomplete, outdated, or presented selectively, faster reading does not fix that. Management conversations, site visits, and independent checks remain essential.
It does not remove accountability. Every conclusion in diligence eventually supports a decision about price, structure, or whether to proceed. Someone has to own those conclusions. Using AI in the process is reasonable; handing it the responsibility is not.
A new question for diligence itself. There is also a newer diligence question: how ready is the target company to use AI after the deal closes? Buyers increasingly want to understand not just current operations, but where AI could realistically improve them and what it would take. That is an operating assessment, not a technology checklist. It asks whether the leadership, processes, and data can actually support change.
The practical view. From my years as COO of a regional M&A firm, the pattern I trust is simple. Use every tool available to understand the business faster and more completely. Then apply experienced judgment to what that understanding means. AI makes the first part stronger. The second part is still the job.