AI is supposed to save hospitality operators time. The question is whether anyone is counting the time spent fixing it when it does not.
Restaurants and cafés are already using AI for customer enquiries, bookings, marketing, rostering and administration. In plenty of cases it will take repetitive work off somebody’s desk. But the saving only matters if that work does not turn up somewhere else in the business.
A booking request is misunderstood. An automated reply gives the wrong opening hours. A promotion goes out for something that has already come off the menu. A customer asks about dietary requirements and a staff member has to step in because nobody wants to rely on the answer they have been given.
Most of those problems will be sorted quickly. Someone makes a phone call, changes the booking or explains the situation to the customer.
Then everyone gets on with the shift. That makes the cost easy to miss.
If AI saves ten minutes answering enquiries but someone on the floor spends 20 minutes sorting out the problems created by those answers, has the business actually saved anything?
It may be worth keeping a tally for a few weeks. Nothing complicated. Note what went wrong, how long it took to fix and whether anything else was involved. A manager may have been pulled away from service. A customer may have been offered a discount. An order may have needed to be remade or somebody may have spent time tracking down what the customer was originally told.
Those small fixes add up. The exercise could also show whether the problem is with AI at all, or simply with the way it has been set up.
If staff keep correcting the same response, perhaps the information behind it is wrong. If a particular type of booking repeatedly causes problems, that step may still need a person involved. If every piece of AI-generated marketing copy is being heavily rewritten before it goes out, the promised time saving deserves another look.
There is another side to this for operators. Staff need to know they can overrule the technology.
Hospitality has never worked neatly enough for every customer interaction to follow a set path. People arrive early, bookings change, allergies need clarification, children need something different and customers occasionally ask for things nobody had anticipated.
Good staff deal with that. AI should make their job easier, not leave them trying to defend an automated answer they can see is wrong.
None of this is a reason to avoid AI. Used well, it could remove a considerable amount of routine work, particularly for smaller businesses where managers are already doing everything from payroll to social media.
But the measurement needs to go beyond how quickly the original task was completed. Count the phone calls. Count the corrections. Count the remakes and refunds. Most importantly, count the staff time.
Then ask whether AI has genuinely taken work out of the business, or just moved it to another part of the shift.
More news here.
