How Racetrack.ai is solving customer bottlenecks for enterprises with AI

What if you as a business owner or manager ended up losing customers because you weren’t able to provide the right information on time? That’s only one of the problems that most businesses face on a daily basis.

Bengaluru-based artificial intelligence (AI) startup FuteMinds Technologies, founded in August 2015 by Subrat Parida, Joydeep Biswas and Shejil Balakrishnan, has developed a set of solutions, which its brands under Racetrack.ai, to generate business insights that claim to tackle most problems around customers. 

“My entrepreneurial ride started from my college days followed by a decade-long journey of successfully building and nurturing two startups. It was during my interaction with clients over the course of my last two ventures that we felt that AI could play a big role in solving most of the concerns that our businesses face in a more efficient and effective manner,” Parida, also the CEO and CTO of the company, told TechCircle.  

Parida, an alumnus of IIM Bangalore, founded his first startup in the area of ERP automation in 2008 and exited the company in 2014. In 2015, he started his second venture which dealt with ecommerce systems.

“I have personally interacted with a lot of academics who were working in this domain and their research raised my interest enough to start researching about AI’s potential on my own. That is when I thought of starting something in conversational AI,” he added.

To take the idea forward, the three founders started developing the company’s first product, a platform called Marvin, that took nearly eight months. For Marvin, Parida said that they had identified that industries were trying to create effective and personalised digital customer experiences.

“However, these experiences were not able to get customers or were finding it difficult. So, personalisation on a mass scale was the core problem that they wanted to solve,” he said.

In order to do so, Marvin, which is an AI-based virtual customer assistant, uses natural language processing (NLP) and deep learning to understand what the customer wants and helps close the sale. 

The initial version of the product was more of a conversational and a lead-generation platform. Later versions were able to pitch on behalf of the client. Also, the product was purely NLP based where the client was able to understand the consumer intent and the context for consumer conversations. 

“Today, we have a combination of multiple technologies where we utilise NLP and NLU (Natural Language Understanding). At the same time, we have a framework where the system learns and understands patterns as well as the sentiments during consumer conversations,” he explained. 

The company signed its first customer within a month of developing the product, starting with the real estate sector and soon expanded to industry segments such as hospitality, banking and financial services, healthcare and insurance.

Marvin works in different industry segments. For instance, in real estate, the product conducts customer sentiment and tone detection to predict the next move. “The property is shortlisted based on travel time mapped with points of interest, in addition to site visit booking with add-on site transfer services,” he explained.

In the case of wealth management, Marvin takes the form of an intelligent trading platform with augmented decision models that can create integrated multi-product relationships to develop a wallet allocation strategy. In the insurance sector, it can be used to focus on customer education and suitable policy identification, followed by a call-to-action and executive visit scheduling for physical verification and medical tests. 

In terms of its go-to-market strategy, the company’s initial focus was metropolitan cities. Now  it has started expanding to Tier 2 cities. The company currently has more than 25 clients. These include real estate developers such as Shapoorji Pallonji, Salarpuria Sattva, Mantri, Kolte Patil, KEYA Homes, HM Constructions and hospitality clients such as Sterling Holidays and Jaypee Hotels. For new clients, the company is focusing on the top 12 Indian cities, Parida said.

It is also simultaneously taking its products to overseas markets. “On the international front, we already have a presence in South East Asia, Middle East and Africa and we will look at expanding this further and deeper within the next few quarters,” he said. The company, however, has faced certain challenges while expanding to other countries. “Conversational AI isn’t a complete product yet and Racetrack.ai has had to add modules and specialisations to Marvin in order to make them effective across different industry segments,” he added.

The company has undergone certain changes in its pricing and business model. “We have now made our pricing KPI-based, from previously what used to be based on the number of conversations a chat-bot had. This has brought in transparency for the customer and their trust towards us in terms of understanding the direct business outcome. So KPI-based factors that are typically considered are the number of leads generated or how many sales were converted online,” he explained.

The change in pricing model along with growing demand for digital services has helped the company grow in the last three years. It claims net revenues at $1.4 million for the financial year ended March, 2019.  

“Our valuation for the last year stands at $20 million and the projected annual growth rate for the next 2-3 years is going to be about 3X,” he said.  

The current size of the addressable market for Racetrack.ai is currently estimated at $140 million and Parida expects this to touch $2.3 billion by 2022.

Racetrack.ai currently employs over 70 people spread across Mumbai and Bengaluru. “Out of these, we have a workforce of about 50 odd technical experts who work on the development side. Over the next year, we shall look to double our headcount in order to supplement the rapid growth we are seeing,” Parida said. 

Parida and his co-founders bootstrapped the company for a year and a half and started with $300,000. It went on to raise around $5 million (Rs 34.3 crore) in 2018 in a pre-Series A round from the ex-president of Biocon Murali Krishna and US-based investor Dev Kagan and other high networth individuals. It is now in the market to raise a $20 million round.

While developing Marvin and adding new capabilities to it, the company has developed another product called Turing for analytics. “Turing can mine large volumes of internal and external data to create critical business insights. It empowers businesses to work proactively addressing customer needs and help in business expansion routes,” Parida said, adding that Turing was built on a hybrid model of neural networks and knowledge-based recommendation engines.