We sat down in a virtual interview with Kalpana Behara, Head of Digitization, Uber, who elaborated on how her team has leveraged technology, particularly automation, across Uber’s businesses to solve real-world problems.
“Uber as an application is really easy to use, you hit a button and get the work done. But there is mind-boggling work that sits behind it. I have been in India all my life but have been working on solving global problems. One thing that I have learned is that problems here in India are not as unique as we think. People all around the world want to take the fastest routes and reach their destination as soon as possible,” Behara said.
“Everywhere in the world, the road systems are constantly changing for one reason or the other. So the problem of keeping the maps data fresh and taking into account all these dynamic changes that happen from time to time is not an easy problem to solve at all,” she added.
Uber takes a two-way approach to solve this problem. The company solicits feedback from its customers, mainly drivers, because they are on the ground and are eyes and ears for Uber in the real world. Uber works with the drivers to understand any of the real time changes that might be happening out there. And the second way is through its automated systems.
“AI is the backbone of a lot of problems we solve across businesses within Uber. My team works on synthesising data to train ML models. Because you need human judgement to start training these ML models. We label millions of data points every year and this makes the system efficient, helps in personalisation and improves customer experience,” she said,
These automated systems flag changes because they are constantly tracking patterns. Once flagged, either by an automated system or by the feedback that Uber has got through customers, its teams work on investigating these data points and take appropriate actions.
The way all of this is orchestrated to solve problems real time, ensures that the customers get the fastest and safest route possible.
“Our systems are always trying to find the fastest and safest route. They are optimised to identify that route for a customer and our drivers are trained to trust the technology and accept the suggestions that it gives. Our pricing models are quite sophisticated. It takes into account not just the distance but also the time taken to reach the destination so overall it is a fair system.
But the moment you add humans, money, safety to it. Disputes arise and misalignment of expectations happens from time to time. And when that happens both the rider and the driver can raise a dispute for the price or the cost. Following which the dispute gets routed to our community operations team who investigate and take the action,” she concluded.