Main idea

A professional user does not compare a business tool only with its direct competitors. Consciously or not, they compare it with the best software experiences they have elsewhere — and those come from an entirely different market.

The shift has already been observed once. Slow, unintuitive, non-real-time tools were accepted at work for a long time on the grounds that they were "business" software. Personal applications became fluid and responsive, and it gradually became hard to accept at work what had seemed normal a few years earlier. The professional standard was raised by experiences that came from elsewhere, without any direct competitor having done a thing.

AI is replaying that movement at another level: it does not merely shift perceived fluidity, it shifts the threshold of what is acceptable. Spending a long time searching for information a system could surface, producing by hand what software could prepare, navigating a complex interface for a simple answer: these gestures were normal and are becoming costly to defend.

Why it matters

This means a position can degrade without any signal appearing in competitive intelligence: the product has not moved, neither have the competitors, and it has become unacceptable.

It also changes what has to be observed. The personal usages of professional users, and the products from other categories they frequent, become a source of information about the level of expectation to come — on the same footing as direct competitors.

Nuances and limits

The transfer of expectation is not uniform: it bears more on interaction, latency and the effort demanded than on functional depth, where business software is still judged on its own criteria.

And expectation rises faster than the ability to pay: a raised standard does not come with a raised budget, which moves the problem toward cost trade-offs.

Open questions

• How do you measure an acceptability threshold that is rising, when it shows up neither as an explicit request nor as a complaint?