Ten years ago, when the buzz words of Machine Learning (ML) and Artificial Intelligence (AI) were just murmurs, was when Anuroop Iyengar started his company, Cogknit Semantics. In this timespan, Anuroop contemplated shutting down Cogknit at least three times. But today, they are one of the leading homegrown ML and AI companies providing voice-first and learning & development solutions in India. When the pandemic hit, their demand doubled and they had to ramp up their capabilities. They are on a growth track that Anuroop believes will register a revenue of Rs. 95 crores over the next three years. The company is currently doing around 10 crores of annual revenue bootstrapped and working towards a 35C ARR over the next 18 months (approximately 5M$ ARR) in revenue within the next three years.
Anuroop was an analog signal processing engineer with Broadcomm before he founded Cogknit. How a signals engineer founded an AI start-up is an intriguing question. But Anuroop says this was his natural progression. For him, be it voice, vision, language or behavior, all become signals in their digital form combined with design and mathematics. And once it struck him how magical the possibilities of AI and ML could be, he got completely drawn in.
It was during his volunteering experience as a teacher with Youth for Seva’s Teach India programme that he found his inspiration for Cogknit. As he taught at the government school in Gottigere in Bangalore, he was fascinated by the impact that the intervention created for the school and its students. The school’s pass percentage that hovered around 15% – 20%, dramatically rose to 50%. Anuroop got thinking about how technology can be leveraged to improve learning outcomes. This became the seed for Cogknit and its first endeavour at developing a product for the education sector.
Anuroop says this was the beginning of a rather unreasonable journey. It’s no ordinary task to bootstrap a deep tech company in India, pioneer it and crack this hard market. All this with the support of funds raised from friends and family, without seeking any external funding. Ullas Sathyanarayana, Deepak Kumar, Ebrahim Mookhtiar and Praveen Rodrigues joined Anuroop in his mission to use ‘cog’nition to ‘knit’ meaningful solutions for life. It just goes to show that self-funded ventures can succeed and it is not necessary to access VC or external funds to grow.
They built an excellent product, but the market didn’t exist – at least then. “One of the biggest lessons we learned, and we learnt this the hard way, is that you have to identify a good market before you build that cool product. A good market is one that has a problem that needs to be solved and where users have the willingness to pay. This is a tad harder in the Deep Tech B2b space”, says Anuroop. After four years of struggle and having burnt through reserves, they pivoted. From learning technology for the education sector, they moved to corporate learning. This change of course catapulted their growth. They developed a learning and development AI platform called ICOG which today has 2.5 lakh users. L&T, one of Cogknit’s earliest customers alone has 70,000 users of ICOG. Now with the pandemic changing the dynamics of digital and learning, the demand for learning tech is skyrocketing.
Their next big avenue came with AI-powered Speech Processing in Indian regional languages. Cogknit launched its ‘voice-first’ application, Autovox, which has collected over 25,000 hours of speech data of Indian languages. This is possibly the largest privately collected data of diverse languages, dialects and accents in the country. Autovox is working with multiple leading unicorns of India the EdTech and companies that are tapping the potential of India’s tier II and III cities.
Talking about the massive opportunity that awaits voice-first technologies, Anuroop says, “Language tech will become as important as UPI and Aadhar. It has all the potential to become ubiquitous. There is huge business value around providing video, audio and text format content in Indian languages. We have micro-entrepreneurs in tier II and III cities who can benefit from voice-based commercial transactions. Telemedicine, especially after the pandemic, is gaining momentum and is waiting for someone to bridge the language divide between the doctors and patients. Also, there is a need to make our legal justice system to be made more accessible in Indian languages”.
Anuroop says that despite the growing demand, Indian language speech is still an unsolved problem. And he believes that this problem has to be solved by homegrown Indian companies who will put the privacy interests of the users and the country above everything else. Anuroop says, “Once customers know that the technology is working for them, their next big concern is privacy. At Cogknit, we have always stood for protecting the privacy of the user and retaining trust.”
Through Cogknit’s survival across the valley of death and in helping them find their resilience, their customers played the role of their patrons and sponsors. They also found their intellectual stimuli from the thought ecosystem of the country. They have mentors from the IITs and IISc, IIIT Bangalore and other leading institutes. Another differentiating factor about Cogknit is their home-grown talent. The team of people is not necessarily the toppers of their institutes or great at communicating in English. But their hunger and aptitude to learn became the strength in building the company with great commitment.
Anuroop says that now as a CEO his focus is on the culture of the organization. “A CEO has to live the culture and not just build it”, says Anuroop as he prepares to set the pace for the next phase of vigorous growth for his company.