AI companies will build principles, but it won’t be easy: Truecaller’s Alan Mamedi
Stockholm-based Truecaller has grown, clocking 356 million active users globally across Android and iPhone. Quite an achievement for something that started as a project to try and solve a problem
It seemed akin to a battle, effectively lost. Spam calls and messages on our phones every day are nothing short of an infliction. The problem compounded, as the global mobile user base widened rapidly. When Truecaller started in 2009, with apps for Symbian OS (think Nokia phones) and Windows Mobile, to be followed soon by Apple iPhone and Android, only the tip of a problematic iceberg was visible.
“I think relevance has never been this strong during all our years. That’s something we never thought of, honestly. We just wanted to solve a problem we had back then. It turned out, that the problem is 100 times bigger today than it was 15 years ago,” says Alan Mamedi, co-founder and CEO of Truecaller, while speaking at the Hindustan Times Leadership Summit 2023.
He calls this a “global problem”, further complicated by the arrival of artificial intelligence (AI) that can be used to build an engine capable of making calls to thousands of users, every second. “The reason is that everything has become digitalised, your bank accounts are accessible from your phone. You’re so vulnerable today that this problem will just continue to grow,” Mamedi said.
Stockholm-based Truecaller has grown, clocking 356 million active users globally across Android and iPhone. Quite an achievement for something that started as a project to try and solve a problem. They’re using tools technology provides, to find solutions to modern-day, real-world problems. The next big challenge Mamedi wants to tackle is fraud.
The app is putting fraud prevention tools in place, with help from AI.
“A user probably doesn’t see how we try to stay two steps ahead because a lot of things are in the background. We’re using AI in various ways around our AI identity platform to understand based on various signals, including from the community, to identify spam,” he says.
The smallest of signals matter. If an unknown number is calling a particular user and also called someone who isn’t a common connection, predictive AI can make sense of the reason, based on historical calling data such as time of day.
Mamedi said he appreciates AI’s improved predictions but is conscious of continued improvement. Scammers have equally effective AI tools, modified for their advantage. “When bad actors are also using AI, how do we make robots fight robots? What data input can we give our AI model, to fight the AI model on the other side? That’s how it’ll play out in the future,” Mamedi predicts.
India is one of Truecaller’s biggest markets, accounting for 70% of the company’s annual revenue. Mamedi calls it their “home market”. But India’s importance goes beyond userbase and subscriptions. There is a significant investment in engineering and product teams. Mamedi is bullish about India’s economic growth, though sluggishness plagues many global economies.
An India-first approach has defined feature evolution for Truecaller. Build a feature for India, then refine it for other countries. AI-powered Assistant rolled out over the summer, is a prime example.
India has diversity, with network connectivity, location, smartphone penetration, and also use-cases. “People have very different needs, and that makes testing of new products, more thorough. It’s going to be very solid when it works with all uses that you have in India,” Mamedi says.
Truecaller Assistant, presently available only on Android phones, extends the capabilities of caller identification with a virtual assistant to answer incoming calls. The next layer of a call received by a robot exhibits advanced speech-to-text technology, to illustrate why a caller may have called while giving the user an option to take over.
Any conversation about AI inevitably veers towards the topic of regulation. “I definitely think there should be some regulations around this,” Mamedi makes his idea very clear.
He hopes those working with AI to build tools, will collectively find an answer to this conundrum. “What we have in front of us is probably in the magnitude of a nuclear bomb, but in the digital world,” he says. A worry is a scenario in which regulation in some countries dictates responsibility with data and privacy but puts companies at a disadvantage compared to regions that don’t regulate.
“When it comes to companies, my sense is they will take responsibility, and build principles around AI that are good for everyone. But when it comes to the interest of different nations, that’s when you will have a tricky situation,” Mamedi predicts.