Software decisions, replicability, and defensible product value

In the past, the absence of a software solution could itself be an advantage. AI is making that gap smaller: anyone building, valuing, or buying software has to ask earlier what is actually hard to copy.

The Financial Times has described how Bain uses AI-generated rough replicas of software products in private-equity due diligence. Not as production-ready alternatives, but as a test of the investment thesis: does the value really sit in the product, or in something harder to replicate?

For now, this is an investor tool. But the logic reaches further. For founders, the validation question is shifting. A USP has always mattered. What changes is that a product idea now has to be tested earlier against its replicability. If a function can be prototyped quickly with AI, the mere existence of the software is no longer enough to justify a software company.

What remains necessary are data, distribution, process proximity, speed, or another advantage that does not disappear in the first replica.

The same question is likely to become more relevant in software procurement. A strong demo, a long feature list, and a plausible roadmap become less sufficient when interfaces, standard workflows, and simple business logic can be copied faster.

An AI prototype does not replace a production system. It does not capture permissions, data migration, operations, security, support, or liability. In good enterprise software, that is often where the value sits.

But it can show which parts of a product are more interchangeable than the demo suggests. The decisive question is not whether something can technically be rebuilt. It is whether a company would replace it at acceptable risk.

The stock market does not prove this, but it does provide a signal. Software stocks are diverging sharply in 2026, as investors reassess which providers can retain pricing power in an AI-shaped market.

AI does not make good software worthless. It makes weak differentiation visible faster.

Source: Financial Times on Bain/vibecoding in due diligence: https://www.ft.com/content/e5bac4d1-b1f8-43a4-bd54-b182d5357af0