How it began
After Picksilk became part of Voonik, I went on to run the ethnic category there. It was a big business, with close to 150 million dollars in sales a year.
When you run something that big, data is everything. You want to know what is selling and where it is selling. Why one saree sells and another does not. What is making sales fall. Every one of those answers shapes the next decision, and the direction you take.
But getting those answers was hard. Much harder than it should have been.
The problem
The data sat in many places, and there was a lot of it. To get even a simple answer, data engineers first had to collect it, clean it and load it. Then analysts had to study it and turn it into numbers and charts. Only after all of that could someone like me make a decision.
It was expensive. It needed costly infrastructure and whole teams of experts. And it was slow. From the moment you asked a question to the moment you saw a number, it often took a week.
A week is a long time when the market changes every day. I kept asking myself: why should one simple question need so many people, and so much waiting?
The solution
The answer I kept coming back to was simple. What if anyone could just ask their data a question, in plain English, and get the answer right away? No data expertise needed. No waiting on engineers and analysts.
That is how NEURAi was born, in Bangalore, in 2017. It was a data exploration platform for growing companies. You connected your data sources, typed your question the way you would ask a colleague, and NEURAi gave you the answer.
We had set ourselves hard cost targets, and the data warehouses of that time were simply too expensive to meet them. So we built our own, from scratch, on Apache Spark and massively parallel processing. It was five times cheaper than what AWS and Microsoft offered back then, and twice as fast.
To let people talk to their data in plain language, we built on a research paper from Google on relational search, and used it to turn everyday questions into SQL queries the database could understand.
Here is NEURAi at work, in a demo we recorded in August 2018.
Growth
In May 2018, startAD, the innovation platform of New York University Abu Dhabi, named NEURAi the winning startup of its Venture Launchpad, and we won the Etihad Airways Innovation Kickstarter Award. It came with a grant and the chance to work with Etihad on a product for the aviation industry. NEURAi was also selected for the SAP Co-Innovation Lab.
For a small team from Bangalore, it meant a lot. People who understood technology looked at what we had built, and believed in it.
What I learned
But building something good and selling it are two very different things.
In India at that time, very few people believed a young startup could stand up to tools like Microsoft Power BI. Large companies wanted the safe, familiar name. And I did not yet have much experience selling to big enterprises. I was learning as I went.
The sales did not come fast enough, and in the end we ran out of money. Paytm Money acqui-hired NEURAi, and I joined them, where I went on to lead the stock-trading experience for four million people.
NEURAi taught me that technology is only half of a company. The other half is selling it: knowing who your buyer is, what makes them trust you, and how long it takes them to say yes. A great product does not sell itself, least of all to an enterprise. I have carried that lesson into everything I have built since.
And it all began with a question I kept asking at Voonik: why does it take a week to get one number?