Skip to navigationSkip to content
Rivigo-Oyo-AI
Reuters/Kacper Pempel
At the cutting-edge.
AI GAME

How two Indian startups are using big data to get ahead—and stay there

By Suneera Tandon

Inside a glass-fronted four-storied building in the heart of Gurugram, 150 engineers are continuously tracking a fleet of 2,500 trucks as they criss-cross the country, travelling thousands of kilometres every day.

Since it launched in 2014, Rivigo has built its business by solving complex logistics problems in a country with a huge infrastructure deficit. The company has devised a relay model, wherein a driver works by handing over the truck to another driver at designated pit-stops. The driver then drives the vehicle back to where he came from. This has helped Rivigo bring down the travel time required to move goods significantly. 

Now, Gazal Kalra, the company’s co-founder, wants to take it to the next level. The idea is to “the take on toughest problems in the (logistics) sector using data,” Kalra said in an interview. For this, the company is adding 40 engineers who specialise in artificial intelligence (AI) and machine learning to its team.

About eight kilometres away, at the other end of Gurugram, online budget-hotel chain OYO is trying to use similar technology but for solving vastly different problems. It wants to infuse AI into its systems to ease the guest experience across its network of 75,000 rooms.

Indian companies are still at a nascent stage when it comes to adapting to AI, and only around 300 startups are using the technology. By 2035, Accenture estimates that AI can add over $950 billion to India’s economy.

And analysts rightfully point out that these are very early days. “Startups in India are still looking at solving some very basic and primitive challenges using data and AI,” said Sanchit Vir Gogia, founder at advisory firm Greyhound Research.

But these two early-stage firms provide a glimpse of what the future might look like.

Complex logistics

Rivigo has over 70 pit stops across India, which have helped bring down the delivery time for its shipments by up to 50%.

The challenge for Rivigo is how to retain its business model and improve efficiency as it plans to double its fleet to 5,000 trucks over the next year. Kalra’s plan, therefore, is to deeply understand all the data the company has collected over the last three years and use it to optimise resources and improve productivity.

“For each consignment, we have captured some 2,000 data points,” said 32-year-old Kalra, a graduate of the Indian Institute of Technology Delhi (IIT-Delhi). Each Rivigo truck is rigged with sensors that track fuel usage and monitor the vehicles’ location and coolant levels. “And now multiply that by millions of shipments which gives me so much data to start predicting,” she added.

Amongst the things that Kalra wants is to predict demand, the reliability of the company’s shipments, pricing, and routing, besides managing fuel costs, improving routing, and optimising inventory management.

For instance, if a customer is to use Rivigo for transporting fragile shipments, the company wants to allocate its most reliable handler in the warehouse and then deploy its more reliable driver to manage such shipments. All this to ensure that there is zero breakage and the fastest delivery.

Big data will give Kalra an insight into predicting demand and pricing for its trucks. Rivigo not only has a fleet of over 2,000 trucks running, it also runs an app for small businesses—an Uber for trucks—connecting third-party truck owners. But deploying these trucks needs advance planning in order to ensure its fleet is used to its fullest capacity.

Now armed with data, the company can predetermine demand for its trucks depending on, say, agricultural seasons or festivals. “I can predict sitting here which client will give an order on which date and I can have my capacity planned accordingly,” Kalra said.

Rivigo is also using data to study how best different goods can be best clubbed in a single shipment. This study of “loading patterns” will help it place goods in the most efficient manner in a truck. For instance, a truck can then carry multiple goods, including fragile ones, without causing any damage, and ensuring maximum products can be transported together.

Another important component of Rivigo’s business is fuel costs. For this, it has a designated team of engineers that closely tracks trucks on the move—monitoring their mileage, assessing the exact time at which they need to refuel, and even tracking how different drivers drive a particular vehicle. This, according to Kalra, can help with better allocation of vehicles on certain routes, and even prevent fuel pilferage.

To make all this happen, the company is building a dedicated AI team of 40, which will also include Vibhanshu Abhishek, professor of artificial intelligence at Carnegie Melon University. Abhishek, a graduate of IIT-Kanpur and Wharton, will begin his work as an advisor in the summer of 2018.

Rivigo is making a big bet on AI, but there is a realisation that this is just the beginning. “We are really picking up every cost we have and seeing what data we have to improve it with the help of algorithms,” Kalra added. “Because, simply put, what you can measure is what you can control.”

Booking made easy

In 2013, OYO launched as a hotel aggregator with a single property. Today, the Softbank-backed company offers 75,000 rooms on its platform across 200 Indian cities.

Last year, the Gurugramheadquartered company pivoted to a franchise model. This means that OYO will no longer only aggregate the hotels on its platform but also operate some of them. And the company now plans to add approximately 10,000 rooms to its platform every month.

This change in the model, and the consequential jump in volumes, means it needs to control its properties much more closely than it has in the past. And this is where technology will help it manage operations better.

OYO already relies on a bunch of in-house apps for both its platform users, as well as its business and sales staff. ORBIS helps its business managers sign on new properties to its network, and Krypton aids its network of 300 cluster managers manage properties, check quality-issues, and address consumer queries.

But Anil Goel, the company’s chief technology officer, has ideas for the future that might look right out of the Jetsons TV show. Imagine walking into your next hotel and not having to check-in, or having robots or drones monitor remotely if your room is clean.

The company is in early stages of experimenting with image-recognition technology that can have multiple use cases.

For instance, explains Goel, the company can have cameras at the hotel that can do an image analysis to identify if there are potential sources of issues such as cleanliness and security. OYO can potentially even use image recognition for check-ins. “Our future vision is that guest should never have to go to the front desk; they should be able to just walk into their room.”

All guests in Indian hotels, in accordance with government laws, are currently mandated to check-in on physical registers. But OYO is piloting a program with several state governments that will replace a physical register with a digital one.

That apart, OYO has been tracking and studying user data for some years now. This helps it customise bookings, go through user preferences, and offer rooms to match what customers actually like. For example, the company can keep a tab on the preferred breakfast timings for its customers and serve them accordingly on each booking. Similarly, user preferences for room types, such as those with a window or a balcony, are offered during every booking.

To make all this happen, OYO will grow its technology team to 400, up from the current 250, by the end of the year.

Armed with the country’s top tech talent and deep pockets, Rivigo and OYO are betting that big data will show them the way forward. Because, in India’s increasingly crowded startup ecosystem, the usual tricks and tools will only get you that far.