We recently sat down with fmr Tinder CPO, Ravi Mehta, to discuss personalization in product with artificial intelligence (AI). Not only did the episode delve into how AI is improving the algorithms on apps like Tinder, but it covers concerns of trust and privacy as it pertains to AI.
Fmr Tinder CPO, Ravi Mehta, and Product Talk host, Nikki Ahmadi, talked about AI products and how they can bring incredible levels of personalization to the products we’re building. You can listen to the full episode of Product Talk above. A few highlights of the episode are available to you below.
On how AI products have evolved over his career
The possibilities of AI products are only going to continue to grow as the technology improves and the barrier of entry is lowered.
“The scope of problems that you can solve is much wider. The quality of results is significantly better. And today in AI, we’re doing things that I didn’t think would be possible in our lifetimes. Self-driving cars, facial recognition, content moderation, video moderation. So, all of those things have really come into play. Because the scope of AI problems has fundamentally changed. We now have the processing horsepower to solve those problems.”
On how Tinder uses AI
It’s always helpful to understand the real-life applications of AI in products. Especially when it’s the favorite product of our founder, SC Moatti.
“There are multiple places within the system that Tinder uses AI today. The first one, which I think people are very familiar with is the recommendation algorithm. So, when you open up Tinder as a single user, Tinder has to figure out who are the people to show you so out of the entire universe of people that are on the app right now. Who are the people that you’re most likely to swipe right on? Tinder’s recommendation algorithm goes through a process and measures a number of different factors to make sure that you’re seeing people who you’re likely to swipe right on and who are also likely to swipe right on you. So that for each swipe that you do within the system, you’re likely to get a really positive social outcome. And the API for that has continually gotten better over time as the number of users has increased, as well as the number of signals that we have about users has increased. “
“Other really interesting ways that Tinder is using AI is in content moderation and safety. So, these are two worlds where we’re combining uses of AI to provide a better, safer, less toxic experience for people on the platform.”
Personalization in product is possible when these things are achieved
AI can make our products feel more personal. Here’s how:
“I think one of the things that we’re seeing today is that the amount of information that users are sharing is rapidly increasing. And that, for many companies, is opening up opportunities but also potentially creating risks are creating challenges in terms of how users perceive a company is using that data. There are two things that are really important in terms of building trust with the user from the perspective of how their data is being used.
The first one is, are you using that data in a way that improves the user’s experience? One thing I’ve seen, especially in the last three or four years with conversations with younger users, is that people do understand that they’re sharing a lot, they do understand that these companies are going to use that data. Now with voice assistance, people know that their voices are being recorded. And so there is this overall understanding that there is, you know, potentially less privacy than there was before. But there’s an implied contract there.
People are okay with that exchange, as long as they’re getting value from it. So, if ultimately the sharing of information from a user to a company results in that user having a better experience, then they feel good about it. And if they don’t value the thing that they’re getting in return, they can walk away from the product. I think the first thing where companies get into trouble is by using that data in a way that benefits the company but doesn’t benefit the user. And those are cases where users do feel violated.”
On the value products deliver with AI
Great product managers know what kind of value their products are delivering to customers.
“It’s really up to app developers to figure out what is the thing within our app is delivering value. It could be social value, entertainment value, other types of value to the user. Then, build a subscription product or microtransaction products based on that value.”
On advice for product leaders
Customer empathy is one of the most important traits of a great product leader.
“Product management is empathy. And I think that’s really important for us as tech leaders and product leaders. Ultimately, the products that we’re creating are having a pretty profound impact on people’s lives. How they communicate what they think, the news that they see, how they relax. All of those things are, you know, firmly in the hands of tech leaders.
It’s really important that we not just build to increase the short term results that our company is trying to optimize for, but build from a very empathetic standpoint so that we’re creating products that are truly in long-termly, durably, valuable to people, and that ultimately improve their lives. So that’s the thing I would leave people with. It’s really our responsibility to do that. And by doing that, I think it’s both the right thing for our users as well as the right thing for the industry and the businesses that we’re in.”