How Lymbyc Solutions is banking on its virtual analyst Leni for Big Data play

How Lymbyc Solutions is banking on its virtual analyst Leni for Big Data play
Photo Credit: Photo Credit: Pixabay
4 Apr, 2019

When a leading e-commerce company in India wants to know what products have been selling fastest on its platform over a 24-hour period, it turns to Leni, a virtual analyst powered by artificial intelligence (AI). But that’s not all what the tool can do. It can even predict what kind of products are most likely to top sales over the next three months.

Leni, developed by Bengaluru-based Lymbyc Solutions Pvt. Ltd, was a product of founder Satyakam Mohanty’s quest to create an analytics tool that could read not just structured data such as numbers but also unstructured data such as text, photos and videos, and deliver meaningful patterns and predictions in real time.

Cuttack-born Mohanty, who comes from a family of engineers and government officers, founded Lymbyc in 2012 with Ashish Rishi. The two worked together briefly at Gurugram-based business process outsourcing firm Genpact. “He (Rishi) worked with me for a short time before moving on to a different part of the business. Even though we didn’t work with each other for long at the start, there was a mutual sense of recognition of the need to challenge the status quo so that things can be done in a better manner,” Mohanty said in an interview with TechCircle.

Mohanty, who graduated in zoology from Bhubaneshwar’s Utkal University and completed his post-graduation in market research from Ahmedabad’s MICA, started as a research analyst at GE’s back-office operations in Gurugram in 2002. When the back-office became Genpact in 2004 after being sold to private equity investors by GE, Mohanty stayed on and rose to vice president.

But it wasn’t enough. So, in 2012, Mohanty and Rishi got together to create a data analytics solution, with on-demand data, that would cater to smaller businesses which could not afford the services of large analytics firms.

Lymbyc claims it had earned revenue of $7.3 million until March 2018. Of that, about $2.3 million was generated in 2017-18. Leni is being used across sectors such as healthcare, information technology and fast-moving consumer goods. The company, Mohanty claims, has 20 clients and its applications range from structured customer relationship management data to unstructured market research and call centre data.

What's under the hood

Mohanty dubs Leni the world’s first virtual analyst. Most prevailing analytics models, he explained, aren't well equipped to break down and understand ‘unstructured’ data. The conventional SQL query model is inadequate to cover them. Also, the prevalent consulting-based analytics processes need too much human intervention. This prolongs its application in real time, which is what Big Data demands.

Those two factors, he adds, call for a model that registers unstructured queries and delivers insights in real time, without human intervention.

Leni is powered by a combination of sub-branches of AI to realise its core promise of understanding basic English language data, context and generate actionable insights, all in real time.

“Leni does two things. First, it understands your question, breaks it down and figures out which is the right data amongst the multiple layers of data that are available in the organisation. The second part that Leni does is the core of the solution and unique to Leni across the world. It can answer ‘What’ and ‘Why’ of the insight,” said Mohanty. An example of how this works is the e-commerce company cited earlier.

The core of Leni’s applications can be broken down into a set of discrete steps. The process begins with a Natural Language Processing (NLP) engine that identifies the data elements and the action a user is seeking through commands. A machine learning ensemble model then identifies where the relevant data lies, fetches it and passes it on to an analytics engine. The analytics engine then applies the relevant analytics model to the data, and sends it to the visualisation engine which then decides what the best depiction is for the given data.

The primary intelligence that puts Leni ahead of its competition is the ability to convert English language queries into NoSQL (Not Only SQL) database queries, which refers to the mechanism of storing and retrieving data through means other than the tabular relations used in traditional relational database. Leni does this with the help of Lymbyc’s proprietary database querying language, MIRA, for which it has applied for a patent. Leni has filed three patents and a few more are in the pipeline, says Mohanty.

Funding and competition

With Leni now gaining traction in the market, Mohanty plans to soon start raising funds to expand business and secure strategic partnerships.

The company raised an undisclosed sum in a seed round in 2012 from MaFoi Strategic Consultants. “Big data and predictive analytics will form an integral part of the growth of any organisation. The industry and the dream of Satyakam and Ashish to build a team and add value to clients made it interesting to invest in Lymbyc,” said Latha Rajan, director at MaFoi.

Companies that offer solutions similar to Lymbyc’s Leni include app marketing and analytics company Localytics and Big Data analytics platform DataStax, both based in the US. In India, industrial internet platform startup Altizon Systems, founded in 2013, and retail management and analytics company Capillary Technologies, founded in 2008, are among its rivals.

It also has to contend with technology giants such as Amazon, Google and Oracle, which also offer such solutions through their respective web services and big data capacities.