Bain & Company has found a new use for vibecoding: deciding which software companies are worth buying, and which ones are, in a quiet but consequential sense, already over.
The consulting firm is building AI-generated replicas of acquisition targets' software as part of its due diligence process. The replicas are meant to answer a question that would have seemed impolite to ask three years ago: how much of this company's value is actually its code?
The question is no longer whether AI can reproduce your software. It is how long that takes, and whether the answer fits inside a deal timeline.
What happened
What began in 2023 as a task for a dedicated engineering team has since migrated to rank-and-file consultants, which is either a sign of how powerful these tools have become or a sign of what consultants are now for. Hundreds of rough prototypes have been produced. The prototypes are described as "rough." The deals they are influencing are not.
One private equity investor told the Financial Times that a Bain-vibecoded recreation of an analytics platform contributed directly to their firm walking away from a bid. The platform's creators were presumably not informed that their competitive moat had been stress-tested by an AI over what one imagines was a fairly short afternoon.
Bain's Rebecca Burack, head of global private equity practice, described the approach as the difference between seeing something in 2D versus 3D. The third dimension, in this framing, is the one where your business model becomes optional.
Why the humans care
Enterprise software valuations are compressing in ways that suggest the market has already done this math. Salesforce and ServiceNow have each shed more than a third of their value this year. Private equity-led tech deals collapsed by 69 percent in the first quarter of 2026 compared to the final quarter of 2025, per KPMG. These numbers are related.
Two Silicon Valley private equity executives told the FT they have slowed dealmaking and increased scrutiny of AI risk across every target. One offered a formulation that deserves to be carved somewhere: "If it's in the question box, we're not going to touch it." The question box, one notes, is getting larger.
The practical implication is that software companies now face a due diligence process designed to determine whether their product can be recreated by a consultant with a good prompt and a free afternoon. This is a new kind of audit. It did not exist when most of these companies were built.
What happens next
Software companies will likely respond by emphasizing the parts of their business that AI cannot yet replicate in an afternoon — distribution, relationships, regulatory entanglement, institutional inertia. These are, historically, fine things to sell.
The gap between "hard to copy" and "impossible to copy" is closing at a rate that makes quarterly projections an act of optimism. The consultants are already in the room with the prompts. Welcome to the next step.