AI Part 2: I told you ‘I will be back'

BANGALORE, INDIA: Facebook's acquisition of Virtual Reality start-up Oculus has probably stirred up more not-so-positive reactions than anticipated. While privacy concerns, credibility of crowdsourcing innovation, fate of Oculus DNA etc stir up the vitriol side, what is also probably confusing people (and getting their attention at the same time) is the relevance of virtual ink in our daily pots and digital plots.

Human-machine interaction has come a long way, from the pages of fiction and Spielberg's fantasy to actual, in-your-face, viable concoctions like Google Glass, Speech Jeeves and navigator Siri, Augmented print genius Layar, medical world magic wands like VRcade, Sensics, CyberGlove or AI-supported war scenarios or aircraft design simulations. Well, the list goes on.

So does the string of doubts, dilemmas and apprehensions that this new world of AI brings forth.

We get to speak to someone who has been working in this space attempting to bring innovation story here in India to a new level in terms of AI.

Uniphore, has been passionately attached to this curve of human-machine interaction and has been working on solutions that can allow any machine to understand and respond to natural human speech, thus enabling humans to use the most natural of communication modes, speech, to engage and instruct machines, as the company aspires.

Umesh Sachdev, Co-founder and CEO of Uniphore helps us with a reality-check as we discuss the technology advancements, industry readiness, consumer mindsets, privacy worries, real world bumps, viability issues, breakthrough gaps etc and take a closer peek at VR.

What's your overall observation on trends in augmented reality? So many attempts have caught the world's eye recently like Siri, advanced robotics, intelligent stamps, high-level biometrics etc.?

Today's innovative technologies are primarily driven by two main consumer trends: personalization and mobility. These consumer demands will be met through a host of new technologies rapidly converging to make more meaningful and intuitive interactions with machines.These technologies contribute to the rapidly growing "Internet of Things," in which there will be over a billion machines talking to each other, performing tasks, and making decisions based on predefined guidelines using artificial intelligence.

So what's making our machines smarter, and enabling them to better interact with each other and humans? Augmented reality applications are becoming more common, adding just-in-time information to our physical world.

How? Especially for enterprises?

Google glass is the leader in this right now. Second-level of emerging technologies are the intelligent electronic agents, which use natural language voice commands to instruct machines, specifically mobile phones. This was most popularly launched with Apple's Siri, and then rapidly followed by Android, Microsoft, and others. Uniphore leverages speech recognition to build applications for businesses, enabling them to give a personal touch to customer interactions. Soon retailers will have a Siri-like sales assistant, and banks will have a Siri-like bank assistant, across voice and data interfaces. Third, combined with automated voice assistants and artificial intelligence, advanced automation and robotics are finally leaping forward after decades of slow growth. With better processing power and bandwidth, robots will increasingly help humans achieve their tasks in more productive and efficient ways.

How would you compare India's stage here vs. global progress, especially in terms of ecosystem gaps, large vendor support, platform extensions, IP evolution, device-level barriers etc.?

Over the last five years, India has claimed 65 per cent of all offshore IT work and 43 per cent of offshore business-process work. Thinking of ecosystem and large vendor support; more and more startups are being recognized in the global economy and nurtured by over 50 incubators across the country. They facilitate business innovation in India and expose the enterprises to global markets. IIT and RTBI helped Uniphore to leverage this opportunity and reach beyond geo-boundaries and gain access to global vendors. The IP Evolution is also happening and we'd be very happy to see it soon. Coming to device-level barriers, India is largely dominated by feature phones today (about 80 per cent of phones in India are basic feature phones), which are incapable of downloading apps or handling local languages. Uniphore's speech-based solutions in local languages help solve these problems. However, this trend is changing now; global players are looking at India as a potential market with lots of opportunities.

How do you stand on the many debates pouring out of ideas like Google Glass, Layar, Siri?

The more human-like technology is, the more people are likely to use it. All these technologies are a step in that direction. While there are many debates about privacy when it comes to these technologies, all of these arguments have happened before. They happened 10 years ago when camera phones were being introduced to the world. Various organizations, places of business and even municipalities tried to ban camera phones. How did that work out? After everyone owned one, the majority of concerns went away. The same will be true for these new technologies. When something brings more value to people's lives than harm, people will continue to use it, and regulatory and protective systems are built around it.

We have come a long way but from where we stand, what apps stay unexplored when we think of leveraging human-machine interaction at entirely new levels?

Human-machine interaction has a long way to go still. For example, Uniphore's speech recognition system is helpful to automate conversations between people and systems. However, our speech recognition engine is primarily focused on keyword recognition, lacking the ability to extract meaningful insights from the customer based on the context of the conversation and the emotional overtone of the interaction. As the technology develops, Emotions Analytics and Natural Language Processing (NLP) will make their way into our applications, giving business much deeper insight into customer interactions and providing users with a very natural, human-like experience with the application. Uniphore's roadmap is focused on these two key areas and to make these technologies more affordable and accessible.

Why or why not would emotionally intelligent technology be a Frankenstein to watch for?

Emotionally intelligent technology is simply another step that we as humans are taking in our journeyto be more efficient and productive - focusing on our creative abilities rather than rote tasks that can be done by a machine. Like any technology, these advances can be used for good and for bad - for saving the world or destroying it. However, Uniphore is committed to leveraging these technologies to help empower businesses and customers to get the information they need, the services they require, and the experience they enjoy.

How has been the traction for Uniphore's solutions in India? Any highlights of some work you have done for Indian customers?

Since its inception in 2008, Uniphore has grown at an exponential rate. In India, our patented technology supports nearly four million registered end users every month. We serve over 60 enterprises across India, including some of the largest financial services, agribusinesses, and government agencies in the country. Overall, Uniphore is riding the waves of personalization and mobility through our high-impact, speech-based applications - enabling machines to understand humans and deliver services on time, reliably and securely. Uniphore is unique in that our speech-based solutions cater to user across all interfaces, from airplanes to a basic feature phones. Our reach extends to rural areas, where speech recognition helps enterprises to overcome the barriers of illiteracy and low-end phones. To those audiences, Siri or other examples of AR are not an option, but the interface is still the most natural and engaging way enable information exchange.

What next on your dart board?

As the leader in multi-lingual speech recognition and enterprise mobility solutions, our dartboard is filled-in with challenges like geo-language expansion (support for more languages), moving from Speech Recognition towards NLP (Natural Language Processing), and upgrading our product to include new capabilities like sentiment analysis/Emotions Analytics, etc. We are also looking to make inroads into global markets by foraying into the non-English speaking countries, especially Middle East, Africa and South East Asia which are not penetrated by speech recognition technology.

What are some key lessons and experiences worth noting so far for a fast growing player in an emerging industry?

When we were a young startup, one of our key lessons learned is how important it is to focus on profitability early, and always, as it allows you degrees of freedom you don't have when you rely upon other people's money. In order to stay profitable, we have made sure to follow an important strategy: Never shove a product down a customer's throat. Just because you have it, doesn't mean they want it. At Uniphore, we are building cool, new, innovative exciting technologies. But we found out fast that these disruptive products were not easy to sell, especially to risk-averse businesses in India.

Therefore, it was critical to just speak to prospects, consult with them, extract their pain, and then sell the painkiller (before building the product). Following this strategy, Uniphore has been profitable since its second year in operations and growing 300 per cent year-on-year for the last three years.