By Saskia Koopman

Copyright cityam

OpenTable has launched Concierge, its new generative AI assistant, or ‘agent,’ designed to give diners instant insights across more than 60,000 restaurants worldwide.

From menu details and dietary options to opening hours, the assistant aims to streamline the often frustrating process of restaurant research.

But in an era of sweeping automation, is AI really the solution diners need, or does it just add another layer of complexity?

A diner’s ‘smart friend’

According to OpenTable, diners spend an average of 22 minutes researching restaurants before booking, while 27 per cent of Brits have abandoned a reservation because the necessary information was hard to find online.

OpenTable’s Concierge was designed with the intention of filling that gap. Embedded directly into each restaurant profile, the AI answers common questions in real time and, in the future, will even be able to make bookings on behalf of users.

“Today’s diners are extremely savvy, and oftentimes they want to know exactly what to order and whether specific needs can be met before they ever step through the door. Concierge makes that effortless,” Sagar Mehta, chief technical officer of OpenTable, told City AM.

For restaurants, he adds, this could “alleviate the amount of time spent fielding questions that diners can now answer on their own”.

The AI draws on OpenTable’s extensive database of menus, reviews, and descriptions, as well as APIs from OpenAI and Perplexity, creating a feedback loop where restaurants can correct outdated information.

“We can incorporate updates from restaurants directly, like menu changes or event schedules, so answers remain accurate”, Mehta explains.

Personalisation vs practicality

Concierge also seeks to make restaurant discovery more personalised.

Mehta notes that diners might want very specific recommendations depending on context: a child-friendly spot for an early Friday night, a gluten-free date-night dinner, or a quick business lunch.

He claims the AI can take these factors into account, offering a tailored list of options with reasoning for each suggestion.

But is AI here solving a real problem, or just adding novelty to a relatively mundane task? While OpenTable touts efficiency and personalisation, questions remain about accuracy, bias, and the risk of ‘hallucinated’ information.

OpenTable emphasises that its answers are grounded in verified data, but in practice, there are risks of AI misinterpreting menus, making recommendations that don’t match real-world availability, or double booking by accident.

The AI arms race in hospitality

OpenTable is far from alone in exploring AI for the dining and travel industries.

Google recently added ‘agentic’ capabilities to its ‘search AI’, allowing users to find restaurants and view real-time availability across multiple booking platforms, including OpenTable.

Meanwhile, Airbnb is moving toward becoming an ‘AI-first application’, using autonomous agents to manage bookings, plan trips, and suggest itineraries.

Both examples show a broader trend of AI moving from novelty to necessity in customer-facing services.

However, they also illustrate the tension between automation and human judgment; Airbnb’s chief executive, Brian Chesky, recently described the need for extreme accuracy, noting that “you cannot have a high hallucination rate” when customers need reliable assistance for bookings and cancellations.

OpenTable is aware of such concerns. Mehta told City AM that Concierge is only the first step, focusing on restaurant discovery and basic queries, with more sophisticated personalisation and booking capabilities planned over time.

“The more you use OpenTable, the better it gets at understanding your needs and adapting to you”, he says.

Ultimately, OpenTable’s AI Concierge is an ambitious attempt to merge restaurant research, discovery, and booking into a single interface.

For diners, it promises speed and convenience, while for restaurants, it offers operational efficiency and potentially more bookings.

OpenTable’s use of AI could end up being a high-tech solution to a task traditionally guided by personal choice or local knowledge, something diners may prefer to manage themselves.

Or, it could be the weight off your shoulders when you’re trying to find a booking for a 20-person birthday, on a terrace, on a bank holiday – which sounds rather useful.