Our AI-driven solutions will help millions of people: Wadhwani AI CEO

Our mission is to apply AI in practical, real-world scenarios and integrate it into existing systems, says Sivasubramaniam

The world is talking about artificial intelligence (AI) but little has been discussed about the technology’s impact on society. SHEKAR SIVASUBRAMANIAM, chief executive officer (CEO) of Wadhwani AI, believes people are yet to discover AI’s ability to solve societal problems. Sivasubramaniam, in a video interview with Shivani Shinde, said his company was the “only AI organisation with a direct presence inside government ministries”. Edited excerpts:

 

How do you define Wadhwani AI’s mission? 

Wadhwani AI is focused on AI applications, and we started our journey in 2018. The first three years were primarily research and academia-driven. When I joined five years ago, my focus was to engage with the marketplace, identify AI-driven problems, develop solutions and deliver them at scale. 

Our mission is to apply AI in practical, real-world scenarios and integrate it into existing systems. To achieve scale, we adhere to key governing principles. The most important is that for AI solutions to be impactful, collaboration with central and state governments is essential. They are the only institutions with the necessary reach. 

Despite the AI hype, we consider ourselves one of the most under-hyped organisations in the field. There’s too much noise around AI, and expectations can often be misleading. Our focus is on three core areas: Agriculture, health care and education. That said, we are open to working on any project that benefits society. 

What solutions Wadhwani AI has developed? 

One of our key solutions is in health care. We have developed a system that scans millions of published articles in 10-12 Indian languages to extract details such as disease names, locations, and timestamps. This data is then provided to the central government to help assess potential outbreaks. We have been running this system for the past two years. 

We are also using a similar AI-driven web-crawling system for the Pradhan Mantri Rashtriya Bal Puraskar awards. The system scans vast amounts of online content and identifies children with exceptional achievements. Last year, we curated a list of over 4,000 names, short listing 500 children who had performed acts of bravery. Eventually, four of them received the award. Without AI, these children might not have been recognised.  

With minor modifications, these solutions can be adapted for other areas such as agriculture, women's safety and more.  

How do you measure impact? 

The impact of AI is measured differently compared to traditional systems. Since AI models need continuous retraining as they are deployed, the first metric of success is adoption: How many people are actually using the solution? 

The second and more critical measure is its tangible effect. How many lives are saved? How many children improve their reading abilities? How effectively can we predict and control disease outbreaks? These are the ultimate goals we set for ourselves.  

How many people use Wadhwani AI’s solutions? 

Some of our solutions have already reached several millions of people, but we believe they still need to mature in terms of depth and usage. For instance, we have an AI-powered clinical decision support system that augments data collection and provides assistance for telemedicine through the government's e-Sanjeevani platform. While the system has been rolled out, we are not yet fully satisfied with its level of integration. 

In terms of usage, if we include e-Sanjeevani, our AI solutions reach 80-100 million people annually. Excluding it, our reach is around 10-20 million users across all our solutions. 

What milestones Wadhwani AI aims to cover in five years? 

We aim to develop about 30 AI-driven solutions that will impact 500-1000 million people. As adoption increases, trust in AI solutions will grow, fears will diminish and new opportunities will emerge. We are still in the early stages — barely scratching the surface of AI’s potential to solve societal problems.  

What are the challenges in working with the government? 

The most important aspect of working with any organisation, especially the government, is understanding the problem you are trying to solve. In our case, we are solving for trust. The first principle of trust is proximity: Maintaining early and continuous communication. 

We are the only AI organisation with a direct presence inside government ministries. We have about 30-40 people working within various ministries. I call this proximal dialogue. In terms of challenges, I have to be honest: If we have ever failed, it has been on our end and not the government’s. The government has been extremely supportive.