- Some restaurants are connecting callers to an AI voice service to place DoorDash and Uber Eats orders.
- Voice AI company Kea takes customers’ orders in phone calls and sends them to delivery aggregators.
- Many people now place delivery orders online, but Kea’s founder says restaurants sometimes struggle to handle phone orders.
A new voice AI service means diners can place delivery orders with services like DoorDash and Uber Eats via phone call.
Rather than having to order on an app or website, customers calling restaurants that use a product developed by Kea, a voice AI company, can simply say what they want to order. Kea then sends their order to a delivery aggregator.
“It’s so much better and easier to just say something and have it done versus let me download an app, let me add everything to the cart, let me customize it, let me add my credit card to it, and then press all those buttons,” Adam Ahmad, Kea’s CEO and founder, told Business Insider.
Before the app revolution, restaurants relied on phone calls for delivery orders. In 2023, more than 85% of Domino’s US sales were made digitally. But there are still use cases for ordering by phone rather than using a website or app, Ahmad said.
“They’re maybe driving the kids back from school and it just makes sense to hit that call button,” he said, describing it as a “30-second interaction.”
When customers call to order delivery from a restaurant that uses Kea's service, they'll be linked to the voice AI, which restaurants can choose from a variety of accents. The AI guides them through the transactions, including suggesting add-ons and recommending items they got in previous orders, Ahmad said.
You can hear an example of how the software works on Kea's website.
"It's not like Siri or Alexa where it's kind of stopping you and making you repeat yourself," he said. "It's very fluid in its interaction."
And if the AI has problems understanding the customer – or vice versa – they can be transferred to a Kea human agent who'll complete the order, Ahmad said. About a quarter of calls made to Kea are transferred to its human agents, Ahmad said.
Kea told BI that it had agents in nine countries but that the majority were based in the US.
Kea uses an algorithm developed by Olo to determine which delivery services are available nearby and choose the best one for each customer based on the cost and estimated time of arrival, Ahmad said. The delivery platforms that Kea can refer orders to are DoorDash, Uber Eats, Postmates and Favor, Ahmad said.
"We'll upsell them throughout the call and then at the very end of the call, they have the option to leave a tip for the driver," Ahmad said. "And at that time, an ETA is provided to them and the order gets placed directly into the system."
Customers don't need to have the delivery provider's app to use Kea's service, Ahmad said. They'd still be sent a link to track their driver, he said.
"We're not, of course, the delivery provider – we're sort of just processing the voice order," Ahmad said.
Ahmad incorporated Kea back in 2018, "before this whole large language model craze," he said. The new delivery-service tie-in aside, Kea has operated a so-called "cashier in the cloud" service for years, which takes pickup orders and answers basic questions like opening hours when customers ring restaurants, Ahmad said.
Kea said that the pizza chains it worked with — which it said it couldn't name due to NDAs — were already using the AI delivery function. It plans to offer the service to its other restaurants later this month, it said. Kea takes a percentage of the orders it takes for restaurants.
Chains that Kea works with include Wayback Burger, Newk's Eatery, and California Fish Grill. Kea declined to say how many restaurants it operated in, but said it was "in the hundreds."
"A lot of them are high volume takeout places, pizza places," Ahmad said. "These are folks that have three to five phone lines at their restaurants and they simply cannot answer all the calls at once. It's just way too cumbersome. And so we're really focusing on those brands first, the ones that just have a lot of incoming phone traffic and ultimately helping them with answering every single phone call versus putting people on hold."