Scalable, deep learning NLPBots for enterprises

By Anusha Ashwin

Man-machine interaction is a global trend now and enterprises in all verticals are slowly bidding apps goodbye to uptake Bots to connect and engage customers. As the mantra to success in any vertical is customer engagement and retention, Bots could be the answer to any enterprise’s calling.

Startups across the world and in India see a huge potential in exploiting and developing customer-centric solutions based on Artificial Intelligence (AI), Natural Language Processing (NLP) and Natural Language Generation (NLG).

One of the early companies to exploit these technologies in real-time application in India is Pune-based Light Information Systems. Founded by three tech-nerds Sanjeev Menon, Animesh Samuel and Sanjeev Nair, Light Systems was born out of a common agreement that the future was in man-machine interactions and the next frontier for technology evolution is Artificial Intelligence and NLP.

Casual embarking onto entrepreneurship

A casual reunion of three friends turned out to be bigger than just a meetup. Conversations among Sanjeev Menon, Animesh Samuel and Sanjeev Nair about their saturated career lives and each one’s quest to be their own boss led to the birth of Light Information Systems.

Interacting with co-Founder Sanjeev Nair, a lot of details about Light Information Systems were unearthed. Sanjeev Nair begins with an explanation to the company’s inception, “The story of Light came to light almost in a filmi way. Three friends who had gone their separate ways after college met after many years. Each of us had our own journeys to talk about, and the common thread at the time was that of the urge to do something pathbreaking and future focused. Sanjeev Menon worked as the principal engineer at Netspeak Corporation in the US and had just exited his first telecom product company that was acquired by a large European telecom major. Animesh Samuel was running KAPS International a training company for the IT sector where he had all the leading IT companies in India as clients. And, I was running operations at Asianet Communications, where in addition to the newsroom automation workflow setup, I also coordinated optimization of the news assimilation network from reporters across the country.”

He continued saying, “When the three of us met and while conversing, we all agreed that the future was in man-machine interactions getting better, and the next frontier for technology evolution would be artificial intelligence and NLP. The thought at the time was that three of us work to build a system that could answer any question it was posed with. And that required us to embark on the journey to build a system that could ‘semantically understand’ text.”

Certainly, those conversations were not over the top. The talks translated into registering a company in Pune. The conversation took place in mid-2012. By mid-2014, three of them had built the technology that had a great deal of success in understanding text input. A team of 15 engineers worked day and night to build on 40+ custom algorithms at the core of the system.

The concept of the technology was taken to a few VCs and angel investors, but as there was no erstwhile proof of the concept being thought of anywhere else in the world, it was impossible for the founders to generate sufficient interest and backing for the company.

Meanwhile, on the other side, the founders managed to win awards and recognition for being innovative in technology. “Our application concept for consumers was being seen as a big social development focused app, as it meant that information would be democratized and made available for all, irrespective of time, place, or creed.”

Post 2014: Positive turn of events

From 2014, the founders got busy. They got their algorithms working, and presented their idea on AI at the right forums. “In October 2014, we were selected as part of the first 11 (and the only company in AI) at the Microsoft Ventures and Reliance backed GenNext innovation hub, an accelerator program. March 2015 we were invited to Pioneer summit to speak about NLP. Late 2016, we launched our enterprise product offering NLPBots, powered by the algorithms we had built over the years. By this time, Facebook had announced its bot platform, Google, Microsoft and Apple had released notes about the advantages of AI and the road ahead. In a blink, everyone was talking about bots and virtual assistants. But at Light, we know that bots are just the interface. The true differentiator is the language understanding, reasoning and generation engine that we have built. NLP Bots have semantic understanding capabilities combined with deep learning and NLG (Natural Language Generation), making it customizable and scalable to suit any enterprise. Come to think about it, we probably could see ourselves like the ‘Intel inside’ to all the automation software out there,” Nair details the events and also outlines the scope of NLP.

Light’s NLPBots

NLPBots is the enterprise product from Light. NLPBots integrate with the client’s online product or service; web, intranet or mobile, to give users an always available help feature via an intelligent chat-bot interface. The bots continuously learn from their interactions with users and give businesses the opportunityto analyse and build on product or service features that drive efficiency and engagement.

NLPBots are custom implemented for each enterprise the company works with. Bots are implemented in automating processes in the area of customer care, employee engagement, reseller management, partner engagement, product/ service support. NLPBots are applicable to any enterprise that requires automation of chat/conversation/system-interaction driven processes across large numbers. Light currently provides applications for the BFSI, telecom, logistics, e-commerce and IT sectors.

The company has been in R&D since inception, and Light continues to be on the same path. Light’s enterprise product was introduced in October 2016 and so far, the company has completed three implementations with many others in advanced pilot stages.

“We are working closely with a few global IT enterprises to enable their customers. Today, 99% of the apps are already passé. Users are no longer driven to install new apps on their devices. In fact, we believe that simple intent based bots that do pre-defined tasks alone, will also be irrelevant soon (if not already!). Bots will need NLP and NLG in order to scale with user interactions, something we always understood and built on day one,” indicates Nair.

The beauty about NLP Bots is that it is ubiquitous and it works on semantic understanding and learns with each interaction. The founders have developed deep learning algorithms on their CNN, RNN neural networks. Enterprises that implement Light’s scalable bots are enabled across all media avenues that like Facebook, Telegram, Slack, etc. In addition to this, Light’s technology is also implemented to enable Natural Language Interfaces across enterprise products and intranets.

Funds & revenues

The company was bootstrapped until early 2016. Mid-2016, the founders raised an initial round from NB Ventures, a fund based out of the British Virgin Isles. The company’s revenue comes from custom product implementations for enterprise, wherein Light has different engagement models depending on the area of the business the solution is deployed in. Although, the company has started generating revenues, it is not yet profitable.

Expansion on the cards

Since the company is already in the making of future and advanced technologies, what more could be instore? To answer that, Nair says, “We are just at the very beginning in terms of the impact this technology will bring to enterprise processes, efficiency and cost savings. Technology wise, NLP and Deep Learning will evolve to make man-machine interactions as close to normal human conversations as possible. With image and voice processing already underway, IoT will soon see NLP adopted in a big way, making you probable see people talking to vending machines for tickets, soft drinks, etc. Retail stores will be able to use technology to capture images from their stores to automatically order the next set of products that need refurbishment, thanks to advanced image identification through deep learning. We have a long road ahead, and its exciting alright.”

Light currently operates out of its Pune office and has a marketing presence in the US East Coast. The founders believe in a lean team and they prefer technology to do most of their work, but having said that, Nair and his co-founding friends constantly exploring partnership opportunities globally. Few months back, Light forayed Africa and has laid plans to expand presence in the West Coast of USA by end 2017.