Most of the enterprises dealing with their customers need an efficient and accurate model of communication to resolve their issues or problems. Many get through while dealing with simple queries, it is complex queries that they get stuck on. And this is where, CogniCor comes into the picture with a knowledge-driven approach in establishing one-to-one interactions for efficiency in business processes.
“We combine a cognitive model and deep learning. This offers an advantage to handle more complex queries because there is a reasoning engine that can derive new knowledge from existing ones without any human intervention,” says Sindhu Joseph, Co-founder of CogniCor, who has also done Ph.D in artificial intelligence from the Artificial Intelligence Research Institute.
Launched in 2012 by Rosh Cherian and Sindhu Joseph under the label of a spinoff from the Artificial Intelligence Research Institute from my doctoral thesis. During its first three years, it developed the platform in collaboration with large enterprises such as Telefonica and France Telecom and deployed it in different pilots and trials.
They mainly serve in three industry verticals, Financial (banking, insurance), IT support, and Telecom.
USP of CogniCor
Some of the main vendors in this space are IBM Watson, Amelia from IPSoft, NextIT, Inbenta, Kasisto, and Artificial solutions. There are three different approaches for creating a chatbot or virtual assistant and most VAs (virtual assistants) generally follow one or a combination of these approaches.
“A script-driven approach is one where the entire conversation is pre-configured and the user has to necessarily follow that specific pattern to achieve the desired result. Chat platforms including Amazon’s Lex and IBM’s Watson primarily follow this script-driven approach,” Sindhu says.
“The second approach is data-driven which relies on the availability of a large volume of data and learning associations between the platform and the user intent, or on understanding the true intention of the user,” she adds.
The company has two main strategies, one focuses on partner-driven direct sales while the other focuses on direct sales to top brands.
In the first model, it extends its technological expertise and knowledge repository to corporate business partners, as in the case of Atos and Wipro, and leverage it to create third-party automation solutions. In the other model, the company targets large corporations through direct sales from CogniCor.
Challenges, Growth and Revenue
“The main challenge we faced initially was related to making the product apt for the market and selling the solution during the early stages of the company. The market had not matured with regards to AI-based solutions, nor was it prepared to try them during 2012 to 2015,” says Sindhu. Now after four years, the challenges have shifted largely towards scaling.
In 2017, CogniCor has marked its entry into the US market by partnering with some of the top brands in the insurance and banking space. Today CogniCor is a team of around 40 engineers, in India, Spain and US.
CogniCor was seeded with an investment of around Rs 30 lakh from Telefonica, a Spanish telecom major, as part of its acceleration programme. It has till now raised around Rs 6.5 crore in funding.
“We are growing in Asia and are also expanding into the US market with partners such as Atos and SixThirty,” quips Sindhu. CogniCor wants to be a global leader in driving the change from search driven engagement to conversation driven engagement in coming years.
CogniCor earned around Rs 1.6 crore in revenue in 2016. The company is anticipating a revenue of $1 million in 2017 and $3 million by 2018.
To step into other markets, the firm is now looking forward to raise more funds. Other than India, the firm is looking to expand its market in South East Asia, the US and Europe via the strategic partners.