SAN FRANCISCO (AP) — Online review aggregator Yelp wants to use artificial intelligence to make it easier for users to find information curated by others.
Yelp users have always been able to dive into a treasure trove of 330 million local business reviews, but sometimes they get drowned in a sea of comments from others about restaurants, doctors, plumbers, roofers, and other eateries. This problem is what the new chatbot assistant is designed to solve.
For example, when a user asks Yelp’s new assistant for a good place to get coffee with their dog, the app will display recommendations along with relevant reviews.
“This chatbot can actually understand 500 reviews, whereas a consumer might say, ‘I read the first five reviews and I think that’s good enough,’” said Yelp CEO Jeremy Stoppelman, who co-founded the company 22 years ago.
Analyzing vast amounts of information and explaining it in easy-to-understand summaries is what other leading chatbots already do, including OpenAI’s ChatGPT, Anthropic’s Claude, Perplexity’s Answer Engine, and Google’s AI Overview.
Yelp believes that pointing out reviews with recommendations and conclusions will help its chatbot stand out. The San Francisco-based company decided to create an AI chatbot to provide evidence to support its findings after research found that most consumers were concerned about technology providing them with false information and fabrications.
“People want AI chatbots to be transparent about where they get their data from, and when they do a local search, they want to see reviews alongside their results,” said Craig Saldaña, chief product officer at Yelp. “So we work to ensure that human connections remain front and center while AI handles all the grunt work of establishing those connections.”
Yelp was looking for a boost amid an AI boom that has more than doubled the value of the tech-driven Nasdaq Composite Index, while its stock has stagnated at roughly the same level as at the end of 2022, just after OpenAI released ChatGPT.
Yelp’s business reviews have always been popular among consumers looking for recommendations on where to eat or shop, but the company hasn’t been able to overcome the habit of people turning to Google almost reflexively when searching for almost anything.
By the time Stoppelman and Russell Simmons launched Yelp in 2004, Google was already synonymous with search, but search results for local businesses were often inappropriate or inaccurate.
To fill the void, Google signed a two-year license agreement to provide access to Yelp reviews. But that partnership fell apart when Google began summarizing different information on different topics, such as restaurant recommendations in a particular area, in a way that gave consumers less reason to click on links that take them to other sites.
This phenomenon has had a negative impact on Yelp and other free online services, which derive most of their revenue from advertising. Yelp relies on Google for over 70% of its web traffic in the US
Tensions came to a head after Yelp accused Google of unfairly usurping its business reviews and favoring its own services. These allegations sparked an investigation by the Federal Trade Commission, which ended in a settlement in 2013 that required Google to make some relatively minor changes.
But complaints about Google’s tactics didn’t stop, leading to a US Department of Justice lawsuit that resulted in a 2024 ruling condemning the search engine as an illegal monopoly. But a federal judge last year rejected the government’s request to break up Google, instead ordering less sweeping changes. The decision comes as more people rely on chatbots for information rather than search engines.
Yelp is pursuing its own antitrust lawsuit against Google, with a lawsuit scheduled for May 2028.
As part of its efforts to diversify and grow its $1.5 billion in annual revenue, Yelp has already licensed some of its data to OpenAI for possible use in ChatGPT, while betting that its chatbot’s focus on connecting people will drive more traffic to its service.
“We really think that with this new technology, we can find the needle in the haystack and have a more personalized experience,” Stoppelman said.
Michael Liedke, Associated Press
