‘Start-ups have the opportunity to leverage data’

As India has become a centre of data science for global firms, start-ups are looking at leveraging local data to build solutions to India’s problems. As more entrepreneurs emerge, investors are not far behind. “We meet five to six AI (artificial intelligence) product companies a week,” said MANISH SINGHAL, Founding Partner at Pi Ventures, in an interview with Alnoor Peermohamed

How did you come up with the idea of starting a venture capital fund only to invest in AI and Machine Learning (ML) companies?

In 2015, I invested in a company called Locus. It applies ML to logistics and that got me thinking about the power of data. When that investment happened, there was a penny-drop moment of “why are we not looking at ML and AI as a space to invest in?” So, we did a lot of groundwork, for five-six months, and about a year ago we launched Pi Ventures. We’re a $30-million fund, which invests exclusively in applied cases of ML and AI. Products in the future would be those that are able to make intelligent decisions based on data.

How many companies have you invested in so far?

In this month, our fourth investment might come through.

AI and ML are broad sectors, but are you investing in product companies or those looking at services?

Product companies, I would say. Two of our companies are in health care. One does blood pathology, using the computer vision. It is filing for patents for what it has built out of Bengaluru. It is planning to go commercial in three months.

People say in India there are not too many product companies using AI in a true fashion. Is that true?

We meet five-six AI product companies every week. Recently, Zinnov published a report, according to which India is the thirdlargest centre for AI in the world and there are three or four factors behind this. One is India has data talent. Big companies had set up their data science teams in India. Data in India is cheap, and available. So it makes sense to consider when you see that for a medical company, data is five to six times more expensive to buy in the US than here. The third is, businesses in India are willing to buy from Indian start-ups and that’s a huge change. The fourth factor is not about India alone but the Internet and computational power have dramatically increased over the years.

You have raised about ~200 crore to invest in AI companies. How fast will you look at deploying that?

Our investment period will be three to four years and we will look at investing in 18-20 companies. We will do fewer investments but in depth. We’d like to get in early with pre-series A funding, and our round sizes will be between $300,000 and $400,000.

How have you trained yourselves to separate fads from true innovation in AI?

The standard things you look are a good team and a large market. We look for three additional things. We look at IP (intellectual property) very strongly and when I say IP, I don’t mean patents only, but the kind of algorithms a company has developed and how defensible they are. The second thing we look at is data acquisition, because data is at the heart of an AI engine. How will a company get and use and once they scale up how much will it cost to buy data? The third factor we look at is how deep a company is rooted in business values.