Chief Product Officer, Yelp
Search every verified Craig Saldanha interview, podcast appearance, and on-the-record quote — each transcript cross-checked by AI and human review to confirm speaker identity. Craig Saldanha, Chief Product Officer at Yelp, has discussed the company's approach to artificial intelligence and user trust in several recent appearances. In December 2024, he described two categories of search intent on the platform: objective queries, such as business hours, which he said large language models can answer effectively, and subjective queries, such as whether a restaurant is romantic, which he said require opinion rather than fact. Saldanha stated that for subjective searches, Yelp uses AI to find the best answer but then connects users to a specific human review rather than providing an AI-generated response, adding that when done well it can "feel like finding a needle in a haystack." He also outlined three reasons he believes consumers trust Yelp: the breadth and depth of its content accumulated over 20 years, the authenticity of reviews from actual human beings, and the ability for users to make meaningful connections with other users or service professionals. Saldanha cautioned that for user-generated content platforms, it is important to understand where trust comes from, saying AI can elevate trust if used correctly but can "very quickly erode that consumer trust" if used incorrectly. In a September 2024 podcast appearance, Saldanha described his mental model for Yelp as focusing on consumer trust, making the experience engaging, and elevating services by removing friction. He stated that the company optimizes for connections between local businesses and consumershol, and that its planning process involves both top-down goals from leadership and bottom-up goals from teams based on consumer and business feedback. Saldanha also noted that Yelp has a product ops team that reconciles dependencies across teamshol. In September 2023, he received the Tepper Recent Masters Alumni Achievement Award from Carnegie Mellon University's Tepper School of Business, where he credited his career success to the support of fellow Tepper alumni, including mentors who helped him with presentations and friends who moved into his home to care for his newborn while he moved abroad for a career opportunity.
“I think of search intents as being of two kinds — the first is objective searches: 'How late is this restaurant open now?' AI and especially LLMs are pretty good at answering these types of queries.”
“The second type of query is subjective — 'Is this restaurant romantic?' — the answers to these types of questions are opinion rather than fact.”
“We recognize that on UGC platforms consumers are specifically seeking an authentic human response, so rather than provide them with an AI-generated answer we use AI to find the best possible answer but then connect them with the specific review.”
“Instead of providing that AI-crafted answer, we connect them with the specific review so they can see all of the context of that entire human experience.”
“When we do it well it can really feel like finding a needle in a haystack.”
“One is the breadth and the depth of our content — Yelp's been around for 20 years and over that time we have hundreds of millions of reviews across broad categories and businesses from restaurants to plumbers.”
“The second is the authenticity of those reviews — Yelp reviews come from actual human beings who have experienced a business firsthand.”
“We allow them to make meaningful connections with other people, whether those people are other users or professionals that are offering a service.”
“For user-generated content products it's very important to understand where your trust comes from — because if you use AI in the right way you can elevate that trust and if you use it incorrectly you can very quickly erode that consumer trust.”
“Tenant number one is delighted a two-sided Marketplace which essentially means we acknowledge that we have two different constituencies that have different goals and needs and while we always look for a win-win there's some instances where there's going to be conflict and we have to use extremely high judgment to resol...”
“We want to be diverse and inclusive in our product thinking so we have a very diverse and inclusive set of consumers and a very diverse and inclusive team and we want to be very thoughtful and deliberate about making sure that every feature that we roll out has the broad feel that we want.”
“We think about our OKRs or goals as what we want to achieve and our tenants as how we want to achieve them. The way our planning process works is there's a top-down set of goals from me based on what I want to see Yelp deliver in the next three to five years, and a bottom-up set of goals from the teams based on what co...”
“We have a small product ops team that owns the overall Yelp roadmap and helps reconcile dependencies across teams, moving from about 60% alignment to 100%. They also act as a unifying force across product, engineering, sales, and marketing to ensure projects progress smoothly.”
“My mental model for Yelp was to focus on three things: consumer trust by protecting and expanding content, making the experience enjoyable and engaging, and elevating services by removing friction and providing guarantees like Yelp Guaranteed to give people peace of mind.”
“The metric that we optimize on is connections between local businesses and consumers. We focus on making meaningful connections whether through advertisers or non-advertisers, via phone calls, emails, or requests for quotes, because that drives the flywheel of growth.”
Subscribe to our youtube channel to learn more about Product Management @ProductSchoolSanFrancisco.
Subscribe to our YouTube channel to learn more about Product Management @ProductSchoolSanFrancisco.
In this episode of Product Thinking, Craig Saldanha, Chief Product Officer at Yelp, joins Melissa Perri to explore user research, ...
Sign in to search the full transcript archive, filter by topic, and access every quote from Craig Saldanha.
The summary and quote tags on this profile are produced with AI assistance from verified, first-person interview transcripts, then checked by our team to confirm the speaker's identity and the accuracy of every quote. See how we verify →