Observing Formula E races in Paris, New York, and Hong Kong last year, each with different weather conditions and road patterns, made it clear to me that the driver must consider a vast set of data points in order to define their strategy for the race. Similarly, as chief digital officer at Genpact, I know businesses must choose the right data points at the right moment in order to make mission-critical business decisions. Electric racing is a great laboratory for learning how to leverage data for insights in the business world.

Formula E is the world’s first fully electric, international single-seater street racing series, and its work is helping to drive electric car and mobility innovations. Technology, and specifically AI, is central to the race’s magic. As an organization, Formula E is very clear that it wants the technology to trickle down beyond the racetrack. No car can hit the road without data. For me, there’s something remarkable about Formula E’s use of data analysis and judgment. And it’s very similar to what businesses think through every day.


A passion for data is key to a winning race strategy. That is what drew Genpact to become not just a sponsor, but a true partner with the Envision Virgin Racing Team. We believe that the intersection of AI, data science, and deep domain expertise is what can help companies be more “instinctive” to predict and shape outcomes.

Formula E drivers reach a city track having spent hours in a simulator, learning the braking points, how hard to accelerate, and the best way to manage battery energy in various track locations and weather patterns. While the engineers worked to improve performance through simulation, they initially lacked deep insights into the drivers’ relative lap times. Genpact used advanced analytics to reveal detailed insights learned from practice runs that now enables Envision Virgin Racing Team drivers to start every race ready to deliver a winning performance. In addition, analytics can identify which sensors or parts should be replaced preemptively, reducing the risk of data blackouts, poor performance, or cars not finishing the race.

The real work is figuring out what data you need access to, extracting it, wrangling it, structuring it, and translating it into something that adds value. And, perhaps most of all, how to do this at speed.

We also extract insights from alternative sources. By cleaning and analyzing GPS data, we create heat maps of opponents’ driving habits so the team can anticipate other drivers’ moves in each race. We help look at what competitive drivers are doing and come up with different scenarios. We look at information associated with the track itself—satellite images, weather patterns, and more. We help the team go through 10,000 simulations before a race even takes place. This requires analyzing significant amounts of data.


This is very similar to the situation corporations face today. There are tons of data, but most of it is unstructured or “dark data,” and therefore companies don’t have access to it and it’s not adding any material value. The real work is figuring out what data you need access to, extracting it, wrangling it, structuring it, and translating it into something that adds value.  And, perhaps most of all, how to do this at speed.

As with most AI programs, prediction accuracy is not perfect out of the gate—you might get to 80% accuracy—and that’s the whole premise of machine learning. Models learn over time with more data. Companies implementing AI need to have a management mechanism to deal with the other 20%. This bit requires a human in the loop with deep domain expertise to train the model and improve the AI gradually over time. And, it requires a strong governance structure to handle the end-to-end automation, manage upstream and downstream processes, and ensure that AI is implemented ethically.

AI can only predict. But prediction needs to translate into action through judgement. In racing you make decisions in a split second. In business it’s often not that different. AI solves the prediction problem, but experts still need to make judgments on this data and then decide what course of action to take. The human in the loop remains critical.

An “instinctive enterprise” needs to adapt to rapidly changing environments. The challenge for large enterprises is that their ecosystems and clients are also evolving quickly. To help keep up and predict outcomes, we must use data, data science, and automation technology to generate the outcomes we want – from winning a race to managing a large company.

With Genpact as its partner, and AI as its neural wiring, the Envision Virgin Racing team is connecting people, processes, and racing knowledge to make accurate decisions at lightning-fast speed. Check out the Envision Virgin Racing Team and see how we are helping it power through the 2019/20 season. And don’t miss a race!

To hear more from Sanjay Srivastava, see this video.