Uma Nidmarty: So how would you define what a chatbot is?
Ayush Jain: My most layman definition of would be, chatbots are chat and bots. So now bots are anything robotic, anything automation. And then on the chat, I see it as more of an umbrella – it could be just a chat or it could be even a voice assistant or it could be any kind of program that you put in information and get out information.
Animesh Samuel: Its automated conversations. However, just a machine to give an automated conversation or an automated response is not going to suffice. In this era, data is of prime importance and especially unstructured data is very key. So if a machine is able to read and understand text like a human then what happens is that you don’t have to actually look out for a keyword but you can understand semantically what a person is talking. The machine can use both the rules that are set there and when it is trained on a neural network is able to respond proactively, intuitively to the user and the user’s needs.
The best advantage that chatbot has over a human is that machines don’t forget. So any conversation can be taken and can be dealt with that.
Uma Nidmarty: How far along do you believe we have come as relates to the AI component of chatbots and how much further do we have to go? From all our conversations so far we seem to feel that chatbots still come across as a very transactional element. So what is your opinion on that?
Ayush Jain: So what I think is that it’s just the beginning, though in fact the beginning has been in place since the 1960’s. That’s when, in fact historically chatbots came into existence. The pace of that growth has definitely become exponential now in the last two to three years.
However, we are just looking at the tip of the iceberg. Being in technology since the last ten years, we can see how technology has been changing and making our lives better exponentially. We can definitely say that chatbots are here to stay. The AI component that is the primary requirement for the chatbot industry to grow and to survive is going to become better and better.
Animesh Samuel: So when you talk about AI, ever since the 1950’s there have always been these waves about AI. When first it was proposed by Turing that a machine could think and make a decision, it was considered preposterous. To make it possible there were two factors that were not existing earlier but which exist today. One is the kind of computing power, the GPU’s that you need, the teraflux of computing power that you need to process all this data, and to learn to put it on a learning model or a deep learning network. And the second aspect of it is also the collaborative nature of the environment today.
So today people are happy to open API’s, people are happy to collaborate. Like our platform is consumed by the likes of Persistent Systems and Capita etc. where they are using our platform to build solutions for their end clients.
Uma Nidmarty: So given that chatbots today predominantly seem to pervade our mind space as relates to consumer activities. How do you think this translates across industries like manufacturing? Where do you feel the real gold mine is?
Ayush Jain: I have been a big advocate on the power chatbots can bring in areas such as governance. For example, the government today has applications, the government today has websites. But what happens is that all these things require a learning curve for the user. The information could be there, but the user cannot access it. The real beauty of the chatbot is that you can make it you so simple for anyone to understand. So now no one has to tell you to learn the website and understand where this information could be or which tab to put into. So with chatbots, all and all we see a future where these things become more seamless i.e. there is no learning curve, there is no installation etc.
Animesh Samuel: In the manufacturing part or the industry part of it, we are doing use cases like for example for Tatacom where a CXO level person, if he wants to approach a prospect – Typically an analyst would have to go through tons of data online. You know who you are meeting? What’s their background? Their hobbies, where they went to school? What’s their company’s strategic focus? Strengths, weakness etc. So typically on a platform like ours, you are able to get that information real time and the beauty is how it gets dispersed by a chatbot.
In an industry or an enterprise, there is a lot of unstructured data. There is a lot of soft communication on the platform like Skype and also chat about the enterprise outside. Any enterprise will have a social footprint today. People, employees, suppliers etc will be talking about the organises. So there is a lot of data in an unstructured format that’s out there which can be made accessible to the right person via chat and at the right time. So in that sense, the industry has a lot to benefit from making sense out of all this data that is there.
The growth of unstructured data is about ten times more than that of structured data. Also, the unstructured data is in silos.
Uma Nidmarty: So you both have talked about chatbots. What they are? And the tremendous opportunities it presents across multiple areas and applications. What do you think are the top three challenges in actually building and investing in this infrastructure as an IT services company?
Ayush Jain: I think you know as an IT services company the challenges are: one- because chatbots they have no UI. They have no other differentiator. The biggest differentiator is your NLP engine, is your AI. So you know as an IT services company you either have to invest in building your own AI/NLP which is totally very difficult. Or you have to find the right partners. In fact, one of the things like Animesh mentioned like companies like his who are the NLP experts and a company like Persistent or a company you know the other IT services company could partner together. Because as an IT services company to be able to cater to the demand and you got to be able to solve your customer’s problem. That can only happen when you have the right NLP and AI in place for the chatbot to really function. Chatbots do backfire. If the right engine is not in place.
So that’s one thing. The second thing is as an IT services company the other thing that relates equally is that the differentiators are really difficult in this space. Because again apart from your interface or your NLP and stuff you don’t have anything else that you can show. Therefore, what we have found a sweet spot is that more and more IT services company are focusing on certain industry segments through which they can build case studies. Or through which they can build their experience or work with multiple partners and figure out what works best. And third is of course educating the customers. A lot of companies are still in the thinking mode and still not open because they don’t realise the value of the chatbot. And last of course is finding good people. So these are the challenges that I see as an It services company for chatboats.
Uma Nidmarty: How about you Animesh?
Animesh Samuel: As a deep tech company, we do not have any domain knowledge apart from the core tech that we are building. It becomes very imperative for us to find the right partner, the right services company who is able to use our technology and deploy it to their clients.
Uma Nidmarty: So I think we have run out of time. And I really want to thank you both for your time and the excitement that you both show around this new technology.