A clutch of ventures has made some early progress in the rapidly emerging field of artificial intelligence, but runaway commercial success may take some time
While running a digital marketing agency, Neerav Parekh regularly updated his clients on their campaign performance with reports and charts that were carefully put together. However, the clients were quickly snowed under the blizzard of data, and inevitably demanded that account managers personally visit them and take them through these reports. This was a laborious process and, rather than plod through it repeatedly, Parekh, a serial entrepreneur, turned to artificial intelligence (AI), the science of trying to make computers think and act like humans, for a solution.
His product, Phrazor, is aimed at automating the process of interpreting data and communicating insights. Having used Phrazor for his agency, Parekh has now sought to extend the reach of his product.
“I realized its enormous potential to change the way data was understood not just in digital marketing but in every other sphere where data was being presented,” he says. “Every company has to send performance reports to its employees or customers. The focus of our venture is to help companies communicate the insights in their data to their people at scale.” For his 14-month-old firm vPhrase, Parekh has ambitious targets -he eyes companies not just in India but also in the US and targets to be in the rest of Asia and Europe in three years. “There is a huge opportunity both in India and in other countries in the analysis and interpretation of data,” he adds.
An Eye for AI
Using AI to pore through, comprehend and piece together numbers and notes is just one use of this rapidly emerging technology. “AI is not new, but its recent reemergence is thanks to the incredible performance of a type of machine learning called deep learning in some of the core problems like image recognition and speech recognition,” says Navneet Sharma, cofounder and CEO of Snapshopr, a visual intelligence platform.
Despite the early excitement, he reckons that there is some way to go before AI explodes on the main stage. “There is a lot of hype around AI but there is very little amount of intelligence in current AI systems,” contends Sharma. “We are very far from creating really intelligent systems. The AI we are seeing today in different applications is, however, very useful in solving a number of industry-specific problems.”
There are other factors that have aided the evolution of AI. This field is getting a lot of traction because of the ease with which AI can crunch huge amounts of structured and unstructured data; enhancements in software and hardware, especially with the emergence of GPU accelerated computing; and, last but the not the least, rapid growth and development of cloud-based infrastructure. (GPU accelerated computing is the use of a graphics processing unit, or GPU, along with a unit, or GPU, along with a CPU to accelerate scientific, analytics, engineering, consumer and enterprise applications.) Recent advances in AI, particularly in the field of deep learning, have helped too, agree executives at AI startups in India.
Despite sounding high-brow and barely out of research labs, AI has actually been in use -and in the news -for a few years. For ex ample, on mobile phones, digital personal assistants such as Siri, Google Now and Cortona all try to learn from the usage and behaviour of phone owners. Video games such as Far Cry and Call of Duty too lean on this field. In the world of business, retailers such as Target and Amazon have used AI to try to predict consumer behaviour, as have banks, media ventures and music-streaming services.
For startups in India, venture capital investments in AI in the United States provide some indication of the exploding interest in this space. According to data from CB Insights, VCs invested barely $415 million in AI ventures in 2012 and this exploded to $2.38 billion in 2015. On a quarterly basis, VC investments in the first three months of 2016 were to the tune of $602 million (143 deals), compared with $491 million (105 deals) in the previous quarter. Still, this financing was down from $901 million in the third quarter of 2015.
In India, startups have focused their energies primarily on the business-to-business side of things, even if some have tried to reach out to consumers.
Sharma’s Snapshopr, for instance, has gained some initial traction and -after much sweat and toil -funding. “It was not easy to get funding to grow the business; we were rejected by more than 100 investors before we got funding,” says Sharma.
He says Snapshopr, which he cofounded with college mates Debashish Pattnaik and Vivek Gandhi, now works with a dozen commerce and retail firms -to help with search and discovery of their products. He believes AI could be a game changer. “I was very much driven by the idea of using artificial intelligence and its immense potential in a wide range of applications,” he adds.
The Next Revolution
Experts and investors in the field of AI say it will bring about the next revolution after the industrial, computer and internet ones. “The advent of machine learning and AI could be extremely disruptive,” says Sanat Rao, partner of IDG Ventures, a venture capital company. “A lot of trivial, repetitive things will soon be done by software.” While large companies such as IBM, Amazon, Face book and Google will build horizontal platforms around AI (see Giants Rush In), there’s a massive opportunity to build vertical solutions for startups. “There is a really big opportunity in sectors such as healthcare and automobiles for AI,” he adds. “Consider a healthcare AI platform that can learn from thousands and even millions of images and can make more accurate diagnosis as it learns.”
Rohit Kumar Pandey wants to do just that with his venture Sigtuple, a healthcare-focused firm that he cofounded with Tathagato Rai Dastidar and Apurv Anand in 2014. Prior to this, the trio worked at American Express Big Data Labs in Bengaluru from 2012 to 2014, building the unit from scratch. Sigtuple leans on the advances in AI. “There is a growing gap between the number of patients and doctors across the world,” says Pandey. “Although a lot of work is being done to bridge the gap, there is a need for an intelligent, effective and scalable solution.” Sigtuple is building a series of cloud based screening tools to increase the efficiency and outreach of the healthcare industry. “We are building a series of AI-based solutions, which automate the screening process (and partially diagnosis) in various fields of health care,” explains Pandey.
The first product, Shonit, is an automated solution for complete blood count. This tool also enables telehaematology and facilitates the complete blood count analysis of patients living in Tier 2 and Tier 3 cities by senior pathologists or doctors in Tier 1 cities in a few minutes. Shonit consists of a standard light microscope, a cellphone camera and a cloud-based AI platform to run the analysis. It just completed a comprehensive clinical validation study at a laboratory in Bangalore. Meanwhile, Sigtuple is also developing solutions for andrology (semen analysis) and analysis of urine, chest X-ray and retina scan. All of these are cloud-enabled, and agnostic of geography and image-capture devices.
As more entrepreneurs try their hand at building ventures around AI, Vinay Kumar Sankarapu, founder and CEO of Arya.ai, along with cofounder Deekshith Marla, wants to provide them with readymade building blocks to hasten their plans. “Our key focus is on simplification of technology through developer tools,” says Sankarapu. “Everyone wants to use AI, but it is really tough to build and scale. We are creating tools and languages that can simplify complex tasks and enable any developer or enterprise to build a complex system in a very short time and scale to millions of users.“
The Problem Solver
So, rather than create a product for a specific vertical, Arya.ai will be like a horizontal platform for developers to build solutions for multiple industries. “(We want to) make AI simple enough so that any developer can start building their own intelligence within hours,” says Sankarapu.
Arya.ai was created as a solution to a problem while Sankarapu was doing research at IIT-Bombay. “With an intelligent research assistant at hand, I could have done better and more effective research,” he says.
“That is when Deekshith and I mulled over this problem and realised that everyone needs an assistant, no matter what their profession. So, we teamed up and created one.”
They have built on their early work to evolve an AI technology plat form. Sankarapu is c autious despite some early progress and funding for his firm. “The biggest threat is hype and unrealistic extrapolations. At every phase, we would see overexpectations from everyone, followed by a correction,“ he says. “Today, AI is used as a buzzword rather than as a product enabler. AI, by its very nature, calls for doing more substantive technology work with little room for hype.”
Deepti Yenireddy and Naveen Verma agree on the great promise of AI. Best friends at IIT-Madras, they went their separate ways after graduation -Yenireddy to engineering giant Schlumberger and Verma to a career in technology and research. They kept in touch all through and decided to come together to found Skedool, a task automation venture. “Rather than choosing AI as a focus area, it all started with a pain point and we realised we could use this technology to solve that problem, given the recent advances in the field,” says Yenireddy. “We are just lucky to be living in an era where such great work is being done and major breakthroughs are happening.”
Skedool works by learning numerous, repetitive tasks across companies and automating them. “There are so many repetitive tasks that you do everyday at work. Our goal is to automate these tasks and give back your precious time to focus on your core skill set,“ says Yenireddy. According to her, scaling up anything in AI is very challenging, since the process needs constant re-evaluation and an extremely focused and dedicated approach to cover the entire breadth of the problem.
Entrepreneurs such as Yenireddy are enthused by the recent progress in AI. “Today, AI is getting very good at outperforming humans in very specific subtasks,“ she says.“The recent success in Go (when Google’s AI-driven computer AlphaGo beat the human world champion in the ancient and complicated Chinese game of Go) is an example of that.”
AlphaGo’s win has also convinced others such as Sachin Jaiswal, CEO and cofounder of the chatbot Niki, that he is on the right path. “We are focusing on revolutionising conversational commerce by creating a seamless and fun-filled buying experience,” he says. “Currently, Niki resides on Android as an app, but going forward, Niki will be a platformindependent, plugand-play software development kit on any interface.”
Jaiswal, who idolises Apple’s Steve Jobs, and Niki.ai cofounders Nitin Babel, Keshav Prawasi and Shishir Modi, believe AI will soon become commonplace. AI is the next big thing, says Jaiswal. As robotics and artificial intelligence continue to improve, in the future, bots will surely replace most conventional jobs.”
The founders of Niki.ai began with chat, but quickly realised that deployment of chat agents limited the scale of operations and proved costly. “Thus we moved to a chatbot, powered by AI,” says Jaiswal. “Our service offerings are bill payments, cab bookings, mobile recharge, home services and food orders.”
What’s more, the company is venturing into travel with hotel and bus bookings this month, along with 25-plus services in the pipeline for the next quarter, including groceries, movie tickets, flight bookings and healthcare.
This kind of ambition clearly has impressed investors -the firm has snared funding from Ratan Tata and Ronnie Screwvala’s Unilazer Ventures. Jaiswal thinks his company can do much more; in the near future, he claims, Niki will become a sales assistant who knows every brand out there, and knows personal preferences as well.
As the field of artificial intelligence continues to develop and business models and investor backing get firmed up, a breakthrough from an AI-driven Indian startup may be on us sooner than we think.
Source: The Economic Times (Delhi)