This Kolkata-based startup is transforming India’s conversational AI space

“Conversational AI has incredible usage in improving customer service and the overall customer experience in the banking sector,” says Tapan Barman, CEO and co-founder of vernacular voice interface platform Mihup Communications Pvt. Ltd.

True to its abbreviation i.e. ‘May I Help You Please’ (Mihup), the Kolkata-based startup is helping global firms deal with country’s linguistic diversity. It has created three Voice AI products which are built on company’s proprietary platforms of Speech to Text; Natural Language Processing; Dialogue Manager; and Text to Speech and cater mainly to the contact centre and automotive space.

The platform integrates Hindi, Hinglish and a mix of other regional languages like Tamil, Bengali, Telugu etc facilitating smoother human-machine interaction. Trusted by multitude of brands including Fortune 500 companies, Mihup has, so far, handled over 100 million customer interactions in India. With its unique offering, Mihup is empowering users to extract actionable and accurate insights that can improve sales, collection and customer services. Besides, it can seamlessly work in low-connectivity areas and even offline.

Mr Barman tells The Plunge Daily about Mihup’s journey so far and the platform’s role in transforming India’s conversational AI space

Q: It’s been five years since you launched this platform. Please take us through your journey.

A: My journey with Mihup started in 2016 with a vision that I had for a long time. I wanted to build a platform that would allow people to talk to machines with the same ease that one has while conversing with another human. My friend and now the COO of Mihup, shared my vision and once we were ready with our plans, we started looking for funds and came across Sandipan Chattopadhyay, the third member of Mihup’s founding team. Together, we incorporated Mihup in 2016.

With our path-breaking innovations in conversational AI, we have managed to carve a position for Mihup in a market which is usually dominated by global giants. Today, Mihup offers multiple solutions such as the VIA (Virtual Interaction Analyst) that can analyse 100% of all inbound/outbound customer interactions of a business to generate actionable insights; AVA – Call Center (Automated Virtual Agent) capable of facilitating automated and precise handling of customer queries through a human sounding AI Agent, and AVA – Auto (a multi-language vehicle voice AI assistant that can perform in-vehicle functions such as car control, media and entertainment etc. Recently we have launched the second generation of our highly acclaimed Virtual Interaction Analyst (VIA) platform in India. With Mihup VIA 2.0, companies get an advanced AI powered interaction analytics platform that can analyse 100% conversations across voice, chat or email channels thereby enabling better Audit and Compliance as well as automation of Quality Analysis. The advanced and multilingual platform can automatically evaluate interactions on various parameters including empathy, product information and accuracy of response provided by the agents.

As a brand, we have come a long way and developed the best-in-class capabilities in Hindi as well as various other regional Indian languages such as Hindi, Bengali, Tamil mixed with English. We are creating market leading innovations in analysis of regional languages and their dialects to create solutions which are ideal to cater to the massive vernacular customer base across India and eventually abroad

Q: Banks are gradually recognising the growing importance of AI. When it comes to customer experience, fintech platforms are doing way better than traditional banks. So, how can banks make the best use of conversational AI to improve customer experience?

A: Conversational AI has incredible usage in improving customer service and the overall customer experience in the banking sector. Traditional banks are mostly lagging behind the private sector ones in the area of consumer experience primarily due to the technological agility and innovation displayed by the latter. A lot of new banks have already opted for automation and usage of AI to better their customer experience. By using conversational AI, a bank can use the humongous data available across the organization and dynamically monitor the entire customer journey. Conversational AI can target banking customers with hyper-relevant recommendations, similar to what we already see in e-commerce websites. Further, banks can also ensure consistency of customer experience throughout their platforms and touchpoints. The quality of the data fed by the banks can directly benefit the outcome of the AI processes and enhances the system’s knowledge. Today, retail banks have almost limitless volumes of valuable customer data. By exploring it through conversational AI deployment, banks can now overcome challenges faced in customer retention, cross-selling as well as improving the customer satisfaction.

Q: Mihup has one of the most accurate vernacular voice interfaces. Please share with us what makes this product stand out.

A: The Mihup platform enables enterprise developers to create automated voice interfaces with full control over brand identity and privacy. We offer three B2B SaaS products.

Voice AI is the future of technology and rapid advancements have been made in this arena recently. Our product has almost limitless potential of scalability as it is unique with the capability of serving the billion-plus Indian consumers in multiple ways. The technology is platform agnostic and can be integrated into almost any kind of usage such as the smart home devices, entertainment devices, wearables, vehicle voice assistants, corporate customer support platforms as well as marketing/customer support systems in physical outlets. The voice AI abilities offered by Mihup also take interaction analytics to an entirely new level not only with the ability to understand and analyse words and data, but also gain insights into the context and conclusion of a conversation. For instance, the NLP-based platform can understand whether a customer is agitated or satisfied with the product and if it is a complaint or a suggestion. Such deep analytical abilities can create insights that boost lead generation, sales, collection and customer satisfaction. We are already serving a number of clients and handling millions of voice interactions every month.

On the voice bot front, the product is not limited in the kind of use cases it can cater to- its deployment ranges from a passenger to a commercial vehicle to serving as the replacement of human agents in call centres- so everyone from a luxury car driver to a truck driver or an enterprise can use our product in the language of their choice – whether it is English for the young urban crowd, or the vernacular for the greater majority of India.

The key innovative factors associated with Mihup would be: (a) Multilingual & Vernacular – Our product is a transformational technology platform which is offering unmatched innovation in the field of voice AI. Until now, the conversational AI platforms have been based predominantly on US English accent conversations with some development in the field of natural language processing. However, what we are doing is fundamentally altering the conversations by integrating Hindi, Hinglish and other mix of regional languages into the conversation. (b) Voice AI on the Edge – The product doesn’t necessarily need internet connection at all times and can seamlessly work in low-connectivity areas and even offline to make it a superior companion for usage in vehicles and rural areas. Thus, it is going to be a great choice for use cases such as car control, media and entertainment etc. Our solutions are finding increasing usage in media, entertainment, consumer electronics, Internet of Things, retail, ecommerce, automobiles, BFSI and various other sectors. Coupled with a low carbon footprint, the product is built in India to solve for India.

Q: An Accenture Strategy study found that 48% of people want special treatment for being loyal customers. How conversational AI can help businesses retain old customers and find new ones.

A: Customer retention and acquisition is directly impacted by the quality of customer service and value additions offered to loyal customers as the Accenture study found. This is where an advanced conversational AI solution can handle multiple customer conversations simultaneously and reduce the wait time for customers which they typically face while contacting a regular agent for support. Further, there is a growing preference for texting over voice conversations and a bot can seamlessly manage conversations across voice, social media, text, email etc. Through inbuilt sentiment analysis and interaction analytics in general, the voice AI can accurately derive actionable business insights from voice and text based customer conversations. The sentiment analytics provides and predicts the overall and individual sentiment for conversations. This knowledge of promoters and detractors can be an important weapon in an organisation’s arsenal. Interaction Analytics also allows tracking of escalations. Proactive handling of escalations is an important aspect in retaining customers for organisations belonging to any domain.

Q: According to a McKinsey study, technologies such as AI can render USD 1 trillion of incremental value for banks which is massive. What, according to you, are the major gaps in the finance and banking sector in terms of technology that require quick redressal.

A: Finance and retail banking are sectors that thrive on customer service, quality of products and their responsiveness. However, these are also areas where institutions have typically been found wanting. An average 30% is lost in revenue opportunity due to poor sales conversion rates, lost bookings due inefficiency in follow ups and inadequate training. There are often long waiting times at the customer care contact centre, a lot of complex processes and an overall legacy framework that often puts the older and renowned institutions at a position of disadvantage. An average of 10% revenue per agent in money lost due to mismanaged collection interactions, inefficient navigation on fraud calls, high TAT on resolving customer queries, missed escalations from customers.

By inducting a cutting-edge voice AI system, banks and financial institutions can eliminate this problem. They are able to offer responsive, dynamic and empathetic customer support round-the-clock. Another major advantage that the banks and financial institutions can get from the conversational AI systems like Mihup is that they have humongous and valuable data bases which often remain unexplored. NPS scores are today at an average of 48% due to poor customer support. Lack of customer insights causes high TAT on resolving customer queries, low FCR and extending required support on time. Conversational AI can also help monitor 100% of inbound/outbound customer conversations.

Also Read: Dream Sports raises $840mn funding from Falcon Edge, others at $8bn valuation

Q: You have clients from Fortune 500 companies. Please tell us little about your business model as well as USP that gives you edge over your competitors.

A: Mihup has built a multipurpose Conversational AI platform which delivers 100% call centre conversation analysis in multiple languages such as English, Hindi, Bengali and a mixture thereof. Having being built in India and for India, it is able to understand the vernacular conversations with far greater accuracy than any other product in the market. Our system works on very low compute which is 10 to 20 times lower than the industry standards, which allows the enterprise-ready platform to function with minimal compute on native cloud as well as on edge networks. In fact, Mihup’s AVA Auto can work with great efficiency without the internet, thus providing its functionality in zero-network coverage areas too. Such unmatched technical prowess and analytical abilities make Mihup prevail over even much larger global brands in the domain.

Q: In May, you worked with Harman International to provide voice control conversation agents in Tata Motors. How do you see the potential market for voice adoption technology across vehicular platforms in coming years?

A: Our work with Harman International & Tata Motors is a first of its kind development in the Indian auto industry that is made in India, for India. Tata Motors is a beacon of innovation and we are fortunate to have been partners with them in our first AVA auto deployment. Mihup AVA Auto offers a diverse capability for improvement over the earlier voice assistant used by them. It offers online, offline and hybrid connectivity options along with various other unique features and vernacular mixed language support. The system can respond to the users’ voice command with a Wake Up Word. Apart from ‘Hinglish’ language support, we are introducing mixed languages like Tamilish and Benglish and by the end of 2022, we will cover all major Indian languages.

Voice adoption in vehicles is rapidly growing and as per an estimate there are going to be 8 billion voice assistants in use by 2023. About 95% of vehicle users are expected to use voice assistants by 2022. The potential is immense and the global solutions aren’t easy to use for the end-users because of the various languages, accents and dialects in use in India. We want to change the way Indians and eventually people across the world interact with machines. By harnessing the power of our NLP and AI tools, we are going to turn India into a global voice automation powerhouse in the years to come.

Q: We have seen a massive shift towards digitisation in the last one and half years due to the pandemic. Please tell us about the opportunities and challenges that Mihup came across along with covid.

A: With the world turning to technology during the pandemic, Mihup’s speech recognition technology too came in handy in helping organisations understand the sentiment of the citizens. In this regard, our technology was used by organisations like Bloomberg to conduct surveys with the local public – for eg. One of the surveys (please check link) was aimed at understanding Delhi resident’s views on the national lockdown due to the ongoing pandemic. Another survey (please check link) was conducted among Punjab residents to gauge their opinion on opening of liquor stores amid the pandemic.

With more and more enterprises turning to work from home, our virtual interaction analytics solution (VIA) also came in handy as organisations could analyse 100% of their customer conversations with the agents and keep a check on agent behaviour and customer sentiment, even when the entire nation was under lockdown and businesses were disrupted.

Also Read: iZooto announces maiden ESOP Buyback Plan

Q: Please tell us about your fundraise, investors and what, according to you, makes them bullish about you.

A: Mihup is backed by investors like Accel Partners, Ideaspring Capital among others. The company has already raised a total of $10 million in funding over three rounds and our latest funding was in December 2020 where we received $1.5 million in a Series A funding round which saw participation from Rajesh Jain, founder of Netcore, and Jayant Kadambi, founder and CEO of YuMe Networks. We are now planning for the Series B funding round by the end of 2021.

Mihup aims to be a gamechanger in the voice AI domain. We have built an extensive platform capable of processing not only the English language, but a diversity of Indian languages and a mixture thereof. This is an area which poses an incredible challenge for even the biggest names in the world. We have already found great demand for the technologies in various domains including the auto industry.