The State of Indian AI Startups: Onwards and Upwards AI space has been garnering a lot of attention from investors. Around $8 billion has already been poured out as an investment in AI startups from 2013-2022, with $3.24 billion in 2022 alone, across over 1,900 Indian AI startups, according to Nasscom's Generative AI Startup Landscape in India report 2023

Consider this: A whopping 70% of India's 31,000 tech startups have leveraged artificial intelligence (AI) in their business, according to 'Indian tech start-up landscape report 2023' by Nasscom. Of this, 65% of the startups are pursuing generative AI and 30% have identified generative AI use cases.

This is not all. The Indian market for AI is set to grow and is expected to reach USD 71.0 billion by 2027 as per the 'State of AI' report by Deloitte. The report also states that enterprises are working closely with agile startups to bridge the technology gaps and to remain competitive in today's market.

So, while AI presents huge opportunities, the big question is how can one build an AI startup? What key strategies are needed for the same? During a panel discussion on 'All about AI startups, at the Tech and Innovation Summit in Bangalore, organized by Entrepreneur Media, and moderated by Priya Kapoor, Features Editor, Entrepreneur India, startups in the AI space came together and shared their thoughts on the same.

Said Saurabh Gupta, Co-founder & CEO, Verismart.Ai, "For any start up, the first and the most important thing is problem identification. Second is data connection. Any AI model requires a lot of data to be trained in the beginning itself. When you scale, you need a lot of data for their training. Building a smart algorithmical and AI engine which could be trained in less computation and can be trained in a lesser data set is what a lot of AI strives for. There could be two different AI startups solving same problem, but their approach is different.

Agreed Harsh Mehta, Co-Founder and CEO, Awarathon: "It's important to first of all to identify the exact use case of AI. What form of AI one is specifically using for it. In our case, it was a simple whatsapp video being shot. As long as you can clearly identify, that's really the key and that journey can take some years and that's the hardest part.

The Co-founder and group CEO of Betterplace, Pravin Agarwala, was of the view that today, the world is changing very rapidly and there are people with a lot of data and money. So when we start something, we have to see what will become common after one or two years. "This is to ensure that you are not building something that is available in general to everyone."

According to Ankush Sabharwal, Founder & CEO, CoRover, it's important for a company to go talk to users to know the problem. "Most of the startups fail when they come up with a problem and they think they have technology for it. It's a very narrow view. If you are trying to solve societal problems, why will you fail? It's easy to talk to businesses. But it's important for a company to go talk to users to know the problem. That way you know their exact problem for which they are ready to pay and then you can build on it."

$8 billion into AI

AI space has been garnering a lot of attention from investors too. Around $8 billion has already been poured out as an investment in AI startups from 2013-2022, with $3.24 billion in 2022 alone, across over 1,900 Indian AI startups, according to Nasscom's Generative AI Startup Landscape in India report 2023. It clearly shows investors' bullishness into the space. However, the startups see that the funding space is changing.

According to Gupta, while earlier VCs used to categorize AI as industry-specific like AI solving, now they are looking at it as an AI application or a business that can be applied to all the possible verticals. "That point of time when it wasn't a buzzword they were focusing on how technology can be a gamechanger for that industry. Now when a VC looks at it, they look at it technology solving at a scale and being sector agnostic at the same time. Their mindset has changed around looking at AI solutions or a product or a platform."

According to Mehta, in the AI space, Gen AI is the buzzword. Now they are asking fundamental questions-path to profitability. Otherwise there was a mad obsession about toplines. Gen AI and strong focus on profitability is really what has changed in the funding landscape.

To build for good

Along with a lot of buzzwords and investors' attention, the AI startup landscape also comes with its fair share of challenges. Added Mehta, "The level of understanding of AI in various forms at an organizational level is not good. Their expectation of what AI can do is sky high. The biggest learning for us is to talk about the limitations of AI. It requires a lot of training. The time and process is required in order to cater to that level of accuracy. It's a very complicated case."

According to Gupta, both large language models and big data have to work together to make AI more efficient and effective. "When we use a tool, we have to define it if we are using it for good. It's all about how we use this particular tool. In terms of adoption of this in the near market, AI if it saturates at 5% of market share in solving a problem, it's really a tool which has half served the purpose and really not make it for a long time for business. These are some challenges which I foresee in the next half a decade from now where people will build something and ensure it is scalable and adaptable and at the same it is built for good," says Saurabh Gupta, Co-founder & CEO, Verismart.Ai.