Most people know the feeling. You are about to leave the house and you want to know how long the wait is at a busy restaurant, so you ask your phone. Soon that question may not need an answer from you at all. Yelp is embedding AI that can place the call itself, listen to the chatter at the host stand on the other end, work out when a table will be ready, and remember that the customer always wants a booth by the window. The app is doing the errand. It is not asking questions anymore. It is completing them, and that small shift signals something much bigger.
The agentic customer has arrived
For the past two decades, the customer side of a business interaction ended at a screen. People tapped buttons, filled forms, and dialed numbers. Software answered questions. Now software will soon be making decisions. The people Yelp is designing for are not just users. They are principals who employ agents, and those agents expect to deal with other agents.
Call this the arrival of the agentic customer. The customer will increasingly hand over the mundane half of their purchasing journey to an AI that will call your front desk, negotiate with your booking system, and compare your prices against a competitor before you even get the chance to say hello.
This is not science fiction creeping in slowly. It is already here, and it is arriving at the speed of the app that calls a restaurant.
Why this should keep operators up at night
Most business leaders are comfortable thinking about AI as something their company deploys. An AI that writes marketing copy, that triages support tickets, that forecasts demand. That is AI working for the company.
The Yelp moment flips the script. Now the AI is working for the customer, and it is knocking on your door. The people your customer sends will not behave like the people who used to sit on hold. They will not accept a canned 'please hold' message. They will press, compare, and move on the second your system is slow or unclear.
The real shift is not that customers get a smarter tool. It is that the customer's side of the relationship is now staffed by software with its own agenda, and it will treat your business as just another node in a network it can route around.
What actually changes operationally
If an agent can call you tomorrow, then your readiness problem is not theoretical. A reservation line that works for a person may not work for an agent, and nobody will tell you the difference until you have already lost the booking.
Consider a few concrete surfaces. Your phone system has to answer a caller that does not identify as human, that may repeat itself, that may try to get a better deal. Your website has to be readable by software that is not looking for a pretty layout but for structured information it can act on. Your pricing and availability have to be consistent, because an agent comparing you against five alternatives will catch a discrepancy in seconds and will not be forgiving. Your policies around cancellations, refunds, and loyalty points become negotiation points rather than terms and conditions nobody reads.
The agent your customer sends is optimizing for the customer's outcome, often expressed as speed, price, or convenience. If your system makes any of those harder, the agent will abandon you. That is not cruelty. That is the new baseline of competition.
The new tensions you will face
This wave brings real risks, and the smart move is to anticipate them before a crisis forces your hand.
Identity and verification become complicated. When an agent calls claiming to act on a customer's behalf, how do you confirm authorization without creating friction that defeats the purpose? Fraud scales differently now. A bot can hammer your pricing or your reservation queue far faster than a bored customer ever could.
There is also a fairness question. An agent that remembers preferences may offer one customer a loyalty discount while charging another the full price. If customers learn their neighbor got the better deal, they will not blame the market. They will blame you.
And there is the trust calculus on the customer's own side. People are still uneasy about handing over decisions to software. The ones who do will expect it to work flawlessly, and they will remember the one time it did not.
What business leaders should actually do
You do not need to build your own restaurant-calling bot to compete here. You need to get your customer-facing systems agent-ready, and that is largely a matter of operational discipline.
Start by making your public-facing information machine-readable. Clean pricing, structured availability, clear policies. If an agent cannot read it, it cannot sell it.
Then look at the interaction points that matter most. Which of your customer service or booking flows can be handed to software on the customer's side? Make those ones fast, honest, and consistent, because they will be stress-tested first.
Invest in verification and fraud protection early, but design it to be invisible to legitimate agents. The people who send good-faith agents should never feel the friction meant for the bad actors.
Finally, set an internal policy. Someone should own how your company responds to automated agents, from pricing to data sharing to how you handle a bot that asks for a favor a human would not. Ambiguity in that space is a liability waiting to be exploited by someone else's smarter agent.
The answer, plainly
Yes, business leaders should care, and the time to care is before the first agent reaches your front desk.
The question in this title is not really about restaurants or reservations. It is about a fundamental shift in who your customer is. That shift matters because the next wave of competition will not be person against person. It will be your systems against someone else's agent, running at the speed and patience of software. The companies that win will be the ones that build customer-facing systems that work as well for an agent as they do for a human, and who treat that transition as a strategy problem, not an IT afterthought.
The app is already calling on behalf of the customer. Your job is to make sure that when it does, it finds a business worth keeping.
