Glimpses of future mobility: EVs would be the first to become autonomous

Kaaman Agarwal, CTO at MetroRide, an EV ride hailing service start-up talks about how AI is revolutionising mobility and what mobility in 2030 will look like

   
EV ride

Worldwide, there are three key technology-driven disruptive trends driving the future of mobility. These are vehicle electrification, connected and autonomous vehicles, and Mobility-as-a-Service. For now, in India, the mobility market has been witnessing a sea change, and the credit for that goes mainly to the transition to electric vehicles.

With regular forecasts on the state of EV penetration, more and more start-ups are coming up in this space. Even all the mainstream OEMs are now focused on EVs, despite multiple challenges. Having said that, the mobility industry is unleashing a dazzling array of innovations designed for urban roads, including new transport concepts such as EV-as-a-service. For instance, MetroRide, a ride hailing service operates fully on EVs. At the same time, AI powered solutions are also re-shaping the mobility arena.

In a one-on-one interaction with SME Futures, Kaaman Agarwal, CTO at MetroRide talks about how EV-as-a-service in unison with artificial intelligence can revolutionise the mobility and car-ride space in India. He also tells us what mobility would look like in 2030.

Edited Excerpts:

How are data and AI integral to redefining mobility? What’s the scope of this market in India?

AI, to me, is any process being driven without human intervention. Particularly processes where the machine acts without a human asking it to act (remember Will Smith’s iRobot from 2004 J). More than the technology, AI needs to focus on the business problems (use cases) that need to be solved.

Rideshare or mobility is a very operations-heavy industry with many moving parts. Some of these variables are drivers, vehicles, rides, locations, etc. To be able to manage these variables efficiently and effectively, rideshare companies go through a series of process innovations. Many mobility companies enforce these process innovations by deploying a huge operations force on the ground, increasing cost and human errors. This is where AI plays a major role. AI should now be used to “Run the Business” and not just to find growth opportunities.  

Also Read: India- The next emerging superpower in artificial intelligence

How does MetroRide leverage data and AI and how are consumers benefitting from it? 

Last year, when we started building MetroRide, AI had to be an integral part of it. We decided not to force-fit AI. Instead, we wanted every process in this platform to be driven by AI. This is when VIKI, our very own, homegrown AI engine, was born. VIKI is now the central nervous system of MetroRide, helping us to solve the business problems that are relevant to us. Some of the business problems we’ve been trying to solve using VIKI are:

Customer Adoption:

VIKI started as an AI-based WhatsApp/SMS-based chatbot. This chatbot was built for users who either didn’t want to install a “new app” on their phones or sometimes did not have an updated smartphone. Using VIKI, users were now able to give it simple commands, like “book me a ride to Dwarka Sec 21 metro station” or “where is my driver” or “show me my last 5 rides”, right from their SMS or WhatsApp. This helped us improve customer adoption right from the beginning.

Efficient Routing:

MetroRide is a fixed route, shared capacity, ride-hailing service. Identifying the right route and pickup/drop points were the key for our business. This is where VIKI evolved to churn data from various data aggregators, obtained footfall data from Google’s Busyness Index and recommend pickup and drop points around a 5 km radius of any metro station. VIKI also created the most efficient routes joining the pickup/drop points.

Driver to Passenger Matchmaking:

VIKI has a complex driver assignment model running behind the scenes that considers more than 20 parameters (like the driver’s direction, current onboarded rides, average rating, gender, distance etc.) before it finds the most eligible driver for a rider. Using this, we were able to deliver an average wait time of 2:01 minutes! VIKI also prioritizes the safety of women by matching a woman passenger with a woman driver during odd hours.

Fraud Detection:

During our beta launch, we realized that fraud is a big issue in the ride-hailing industry. Fraudulent activities like drivers taking cash rides while on MetroRide duty, asking customers to cancel the ride and pay in cash were the most common and have not been solved by any ride-hailing platform yet. For us, it was important to keep a check on this as it was resulting in huge revenue leakage and incorrect ride assignments.

We decided to build an inexpensive (less than $20) camera inhouse. This tiny camera takes time-lapse images from inside the vehicles, every two minutes, and uploads them to VIKI. VIKI’s image processor reads the image to identify the number of people in the image and matches it to the number of seats booked on the vehicle. This identifies any fraudulent rides at any given time. This image processor also detects drunk or distressed drivers using their facial expressions, enhancing the safety of every ride.

Operational Efficiency:

For us, scalability is one of the most important aspects of our business. Hundreds of vehicles and drivers required close operational monitoring. Hiring a huge on-ground operations force is inversely proportional to scalability.

We needed a mechanism to be on the field, remotely. VIKI’s location intelligence gives us 100 per cent real-time intelligence on the driver’s location and driving patterns, in the comfort of our office while deploying minimal on-ground staff, increasing our operational efficiency.

Also Read: One in two cars sold will have electric powertrain by 2030: Report

Focused Hyperlocal Advertising:

While serving rides, ride-hailing platforms generate a huge amount of data. This data includes the riders’ profiles and commute patterns. We decided to capitalise on this rich data, without selling it to any third party. As the simplicity of the app is our priority, showing advertisements on it was out of the question.

Ours being a hyperlocal phenomenon, we decided to install LED screens inside our vehicles to show hyperlocal advertisements to our riders. VIKI’s hyperlocal profiling engine hand-picked advertisements that were the most relevant to the riders traveling in our vehicles, based on their profiles and commute patterns.

So, if a 25-year-old person is traveling in a vehicle he would most likely see an advertisement for a nearby gym or theatre instead of an advertisement for a preschool. These advertisements were served on the LED screens, giving us a second source of revenue. This helped us keep our rides cheaper and our drivers made more money than what they were making while serving rides outside our platform.

EV-as-a-service, how is this model growing in India? Also, from your point of view as a stakeholder, what areas should this sector work on to get promising growth, barring EV charging infra?

EV-as-a-service is an absolute must for India’s EV Ride adoption. That being said, it is even more important to do it right. EV-as-a-service should not just be about giving EVs on rent, rather it should be a bundled offering. Now is the time to combine all the eco-system partners in a single deal, especially for the ride-hailing platforms like us. This bundled offering should include the vehicle, its charging, its parking, security and, most importantly, data.

Assume its 2030, what does EV mobility in India look like to you then?

The three largest costs of the ride-hailing industry are—fuel cost, driver cost and vehicle utilization.

Fossil fuels are not getting any cheaper, making electric vehicles an obvious choice for the future. AI can help EV adoption in the ride-hailing business by optimising battery consumption and increasing charging speed. About the other two costs (driver cost and vehicle utilisation), AI-enabled autonomous driving is the solution for that.

Self-driving autonomous cars are coming, and they are coming sooner than we think. EVs would be the first to become autonomous. ICE vehicles, if at all, would follow. This would bring about a paradigm shift in EV ownership, bringing in an “Airbnb effect”. Once autonomous cars are mainstream, ride-hailing platforms would “subscribe” to our EVs.

Imagine sitting at home binge-watching a series on Netflix and your EV automatically accepts a ride request, comes out of your garage, serves the ride, and comes back in the garage to charge. All this happens without you even knowing, and at the end of the week, you get a cheque from your ride-hailing platform. This is what “The Future” looks like to me.

Also Read: Hopcharge: EV charging at your doorstep

What’s the current presence of your company and your roadmap for it down the line? 

Currently, MetroRide is serving rides in Hyderabad, Bangalore, Noida, and Delhi. More importantly, MetroRide is forming collaborations with various metro systems now to offer a seamless, multimodal commute to our users. We have already tied up with the Hyderabad Metro to integrate the metro’s ticketing within our platform. Using this, our users will not just book their first and last mile but also their “metro mile” on our app, hence creating a seamless commute from home to work. We are also in talks with various other metro systems to integrate with our platform.