Most restaurants pay a flat monthly fee for their reservation software. It’s predictable, it’s simple, and it’s been the standard model for years. That’s starting to change, and it’s worth understanding why before it affects your next contract renewal.
Across the software industry, AI is driving a shift away from flat subscriptions toward pricing based on usage or outcomes. The logic is straightforward: if AI is doing work that previously required human effort, vendors want to price on the value delivered rather than just access to the platform.
Zendesk was one of the first software companies to make this move. In 2024 they shifted their AI customer service tools to outcome-based pricing, charging per issue resolved by AI rather than a flat per-seat fee. Gartner and Deloitte now predict that by 2030, at least 40% of enterprise SaaS spend will have shifted to usage or outcome-based models.
In hospitality this could look like
- paying per booking generated rather than a flat monthly fee,
- tiered pricing based on how many AI features you actually use, or
- outcome-based models where the vendor takes a percentage of measurable value - revenue uplift, no-shows reduced, covers filled.
None of this is widespread in restaurant tech yet. But as I wrote a couple of weeks ago about the Sequoia thesis, the direction of travel in software is toward selling the work rather than selling the platform. The vendors who move first will market it aggressively.
For operators, the key thing to understand is that these models aren’t better or worse than flat subscriptions - it depends entirely on the numbers. A per-booking fee sounds attractive when you’re quiet but expensive when you’re busy. An outcome-based model sounds fair until you’re paying a premium on revenue you’d have generated anyway.
That last point deserves attention as operators have seen this before. OpenTable ranked higher in Google than the restaurants themselves, became the default booking path, and then charged per cover for bookings the restaurant would have received directly. Outcome-based AI pricing carries the same risk. If a vendor charges you for a “booking generated by AI” or a “no-show prevented by AI,” the question isn’t whether the outcome happened, it’s whether it would have happened without AI. That’s the attribution problem, and most vendors aren’t set up to answer it honestly.
Before your next renewal, ask your provider what their pricing roadmap looks like. If they’re planning to introduce usage or outcome-based elements, you want to know before you sign. Run the numbers on any new model against your actual booking volumes. Vendors will show you the best case so model the worst case yourself. And if outcome-based pricing is on the table, insist on clear attribution methodology - how does the vendor prove that AI caused the result, rather than just being present when it happened?