Your next customer may send an AI

Somebody always ends up booking the table.

Imagine it’s your turn. Six people, Saturday night, a show at eight. One vegetarian. Somewhere close enough to walk, nice enough to feel like an evening out, and affordable enough that nobody has to ask what ‘market price’ means.

The group chat has contributed three restaurant names and very little help. You open a menu, then a map, then a booking calendar. The first place has room at nine. The second looks promising until you notice the menu is from last summer. The third wants you to call.

You can do all this. You would also quite happily let someone else do it.

That is the appeal of asking an AI assistant. And for the businesses waiting at the other end, it raises a question that reaches well beyond search rankings: how much of doing business with you can a customer hand over?

We have been here before

There was a time when a restaurant’s web address was something the owner proudly pointed out on the bottom of the menu. It meant you could look at the dining room from home. Check the hours. Send the link to your sister. The place existed on the internet, and that was an achievement.

Then people began looking for restaurants whose names they didn’t know. A lovely website was useful once somebody arrived; getting a stranger there became a business of its own. Owners learned why the restaurant down the street appeared in Google and theirs did not. Having a website and having anyone visit it turned out to be two different accomplishments.

Meanwhile, the customer with six people to feed was still doing the legwork. Finding the restaurant didn’t reserve the table. At eleven at night, the phone number on the website was an instruction to come back tomorrow.

Online reservations changed that small, maddening detail. Shops did the same with checkout. Suddenly, the moment a person decided could also be the moment they became a customer. The website had become part of how the business operated.

These changes overlapped. Different businesses adopted them at different times, often reluctantly, usually while busy with everything else. But the expectation was hard to reverse. Once people could book a table from the sofa, ‘call us during opening hours’ felt like work someone had handed back to them.

AI enters this history with a familiar attraction: there is more of that work it could take off the customer’s hands.

An evening out, in one request

Go back to the group chat. Instead of translating everybody’s preferences into a series of searches, you give the assistant the whole problem: ‘Find a good Italian restaurant near the theater for six on Saturday. We need vegetarian options and time to eat before the eight o’clock show.’

A person hears the evening you are trying to arrange. Search has traditionally made you break it into pieces. A conversational assistant lets you keep it together, ask follow-up questions and work toward a choice.

For a restaurant owner, that is an intriguing change. Your most persuasive selling point might be something quite specific: a proper vegetarian menu, a short walk to the theater, an early sitting that makes dinner comfortable rather than hurried. Those details are suddenly central to somebody’s decision.

They need to be available for the assistant to use. ‘A dining experience for every occasion’ tells it less than the menu, the address and the booking times. The particulars that make a place right for an evening matter more than another paragraph declaring it exceptional.

Google’s August 2025 AI Mode rollout offered a version of this: searching across restaurant sites and reservation platforms for tables that matched a request, then sending the customer to a booking page to finish.¹ That is a meaningful amount of organizing taken care of before a diner reaches the restaurant’s own site.

You can see why businesses want to appear in those answers. A recommendation can bring you a customer who has never heard of you, because your restaurant fits the evening they have in mind.

Then the customer reads the suggestion and replies, ‘That looks good. Book it.’

Now the restaurant has to answer

Up to this point, the assistant has been helping someone decide. Now it has been asked to make an arrangement. The restaurant’s calendar, booking rules and confirmation process enter the conversation.

Is that table actually available? Can a party of six reserve it online? Is there a deposit? Does the customer have to approve a charge or finish a step personally? What comes back when the booking succeeds? A charming description of the dining room cannot settle any of those questions.

For the owner, this is where the opportunity becomes tangible. An assistant that can bring an authorized request into the restaurant’s existing reservation service offers a way for someone else to handle more of the organizing. The restaurant still sets the terms, accepts the reservation and welcomes the guests.

The route will vary. Sometimes the assistant hands the customer a useful link. Sometimes a supported connection allows it to carry out more of the task. OpenAI’s March 2026 shopping update, for example, emphasized product discovery with purchases completed through merchants’ own checkout experiences.² The business’s existing systems remain essential to what happens next.

That makes ‘Can AI do business with us?’ a practical question to investigate with the services you already use. A booking provider may support a useful connection. A shop may be able to supply current product information. A service business may need a clear way to receive an estimate request. Each has a different job to finish.

And each needs an ending both sides can trust. In our evening-out story, that means six people under the right name in the restaurant’s reservation book, with a time and confirmation the diner can rely on. Otherwise, someone is still arranging dinner.

Make room for the next customer

A business owner could be forgiven for hearing all this and thinking: another thing to fix. You already paid for the website. You already have a booking system. Keeping the menu current should not require a second career.

Those investments are the starting point. The useful work is to see how they serve a customer arriving through AI, then improve the parts that stand in the way. Begin with the kind of request you would love to receive. Follow it far enough to find out what a customer, or their assistant, can actually accomplish.

That is the purpose behind Transact, the main offering from AI Promo Services: helping your business become discoverable, compelling to choose and ready to do business through AI and the commerce, scheduling and payment services you use.

The work starts with your business as it is. What makes it worth choosing? What can an assistant find out about it? Which supported route leads to a booking, purchase or enquiry, and where does that route need attention? From there, the job is to connect or improve what is needed and check that it works.

Sometimes that means clearer website information or a better connection to an existing service. Sometimes the earlier problem is getting discovered at all, and useful articles, local information or social content have work to do. The priorities should follow the opportunity to win a customer.

Think again about the person organizing Saturday night. They do not particularly want to become better at searching restaurant websites. They want to tell their friends where to meet, put the phone away and look forward to dinner.

Your next customer may ask an AI to get them that far. What would it take for the reservation to be yours?