LVInsights AMA: Big money, big data, and how to fish for the right deals in a pool of AI startups

Gone are the days when machines making decisions like human beings seemed like a distant future. Mankind has come a long way from 1950 when Alan Turing first pondered over the question if machines can think.

Now concepts like artificial intelligence and machine learning are mainstream in the startup ecosystem.  Its applications in improving efficiency in sectors such as healthcare, finance, defense, commerce and even gaming are undeniable.

Several new-age businesses are turning towards AI with an aim to tap into the global AI market value which is expected to reach nearly $450 billion this year, according to the International Data Corporation (IDC). 

According to the Stanford AI Index, investors poured $93 billion into AI in 2021 alone. For investors, though, it may look like the time to go big bullish over AI, but remember to proceed with caution.

“Every second startup uses AI in its pitch deck. So, while AI may be a great place to invest in, it is not always easy to do the right investment,” says Rahul Agarwalla, Managing Partner, SenseAI Ventures, during the fifth edition of LVInsights’ Ask Me Anything (AMA) session.

He also goes on to say that a lot of companies “fake it” when labeling themselves as an AI company. So how should investors fish for the right deals in a pool of AI startups?

Variety as the key to deploying AI 

Rahul believes the key is to figure out if there is a real use case for AI. It is important that the businesses have data and its own proprietary model because the user will be different from the company that develops AI. He suggests that a business should not resort to AI just because it can, as data and processing power tends to be expensive. 

In order to deploy AI-based solutions, one of the crucial criteria to keep in mind is variety. Variety of anything can lead to complexities which in turn can be solved through AI.

Citing an example, Rahuls explains  that if there is a startup automating traffic signals, there is only one input which is the number of cars in each direction  and only three outputs which are the colours red, yellow, and green.

“The amount of variety of this use case is very variable, which tells me I would not invest in a startup which wants to automate traffic signals using it. Is there value? Yes. But is the cost justified? I don't think so,” he adds.

On the other hand, self-driving cars are on the other extreme of the spectrum, as it must consider a variety of factors such as time of the day to automate headlights function, the weather to automate features like wipers. Self-driving cars must also understand  the traffic rules and lights and be equipped to detect the dynamic elements such as humans, vehicles, or any other objects, animals among others while on the road.   

Hence many variables are to be considered by self-driving cars  to decide if it needs to turn left or right, speed up or slow down, brake, change gears among others. This  variety makes it an ideal use case for AI. Rahul adds regardless if AI has been able to completely solve for self-driving cars yet, it is a practical use case of AI and has also helped further the technology. 

Data and other frameworks 

To build scalable AI-based solutions, businesses need to ensure availability of high-quality data. Rahul says, “Without great data, you don’t have great AI and availability of data is rapidly increasing.” 

While considering an AI deal, he evaluates if the proprietary algorithms are optimized for solving unique problems and if the team has deep understanding of big data. 

“Investors, please look at the team. If you ignore everything else and get this right, most of the time, you'll make money. That has been my personal experience through my life. You ultimately invest in people and not in companies,” he asserts.

In the early stage, revenue might not be critical as long as there is an early, working prototype of the product powered by strong algorithms and data. 

This is why his artificial intelligence-focussed fund SenseAI Ventures, follows a unique V-DAT (which expands to Variety, Data, Al IP, and Talent) framework when evaluating AI investments.  

SenseAI Ventures has been investing in businesses across sectors such as SaaS (Software-as-a-Service) companies spanning cybersecurity, ecommerce, BFSI, transportation and logistics, B2C (Business-to-Consumer) spaces, healthtech, fintech,  creator economy and most importantly, in the on AI ecosystem. 

“While we invest in companies that have AI as part of their offering, we are also investing in people who are providing the tool to build the entire AI ecosystem. This ecosystem needs to be built and we are at the start of the J curve,” he adds.  

AI: A superwave

Rahul compares the AI-led transformation with a superwave, specifically, The Great Wave off Kanagawa painted by Japanese artist Hokusai. He describes – both AI and the superwave – as a structural shift of long-term in nature with irreversible consequences.

Underneath the anticipation around AI and its impact, the AI ecosystem is constantly growing and evolving with better models and algorithms. For instance, the time to create an image classification model has come down by 94 percent in the last four years.

Rahul is very bullish towards AI’s potential in creating new kinds of business models and products that may change how the world works. He believes that now is the time for long-term investors to get in the game because while the AI boom is international, the tech is even younger in India.

AI’s deep penetration into our daily lives might have already started with us taking Siri’s help to play a song!