AI (Artificial Intelligence) & Machine Learning (ML) have certainly been a booming topic of discussion amongst the startup ecosystem across the world. In most ecosystems and in almost every new age startup, AI / ML have almost become a hygiene parameter.
With this reference, we had set up a panel on our annual flagship event – “Blume Day 2019”, moderated by Arpit Agarwal, who leads Deep-tech investments at Blume, to answer one specific question “Is AI really helping solve real life problems?”.
Our panelists included founders of Blume’s AI /ML firms who are solving core industry problems using these technologies – Nishith from Locus (Supply chain optimization), Abhimanyu from Agara labs (AI solutions for addressing customer interaction problem of enterprises), Rishabh from Belong (Outbound Hiring for Enterprises) & Apurv from SquadRun (Future of Work for Enterprises) spoke at length about the use of AI in solving different problems.
We would like to highlight the three key insights (with a few subpoints in some cases):
(1) Does AI solve a classification problem or an optimization problem?
a. Nishith from Locus says – “AI solves a classification problem for them (for ex: understanding unstructured data from addresses) rather than the core optimization problem”
b. Abhimanyu & Rishabh were of the view that “AI helps them classify human communications and emotions, which in turn helps them with their prediction algorithms”
c. Essentially, AI is a tool, which can be used for classifying real life data points, which in turn makes your optimization problem a bit easier to solve.
(2) Expectation setting to the client is one of the key responsibilities for a founder
a. Apurv highlights three types of customers – (1) Who understand AI & its limitations, (2) Who want to understand AI, (3) Who don’t understand AI
b. Nishith added “50% of my time goes into helping my clients understand that AI is not magic”
(3) Will India drive the “Big Data-ization of the world”? – Do we have good quality data?
a. Basis learning from Squadrun customers in US, Apurv says that enterprises in U.S. have far more quality data than Indian ones, a view which was supported by others
b. However, when it comes to unstructured data in healthcare, security etc., India has proven to be a good training ground for most early stage companies