A customer is the most crucial for any business enterprise. Today, with the evolution of the digital economy, a number of things have changed for the end customer – from buying grocery online to availing loans.
Realising there were processes that can be automated to a great extent to enhance the customer experience, IITian couple Manish Gupta and Rashi Gupta started Rezo.ai in 2017.
The Noida-based startup uses machine learning and NLP to automate enterprise workflows with limited human intervention. It also enables automatic conversational responses to customers.
Manish says, "In the last 15-20 years, both Rashi and I have travelled a lot. We had been in the US and Europe for some time. What we realised during this period was the difference in the customer experience. If you interact with a brand in the US, Europe, or Singapore, the kind of experience you get as a customer is way different from what you get in India."
Today, Rezo.ai has Delhivery, CarDekho, and clothing brand W among its clientele.
Manish and his wife Rashi Gupta are IIT Delhi graduates with a masters in Mathematics and Computing.
The duo later went to the University of Helsinki where Rashi got a PhD in Biometry, and Manish did his research in statistical analysis. Before starting their entrepreneurial journey, Rashi was working with WNS and Manish was the CTO at Rategain.
The couple was working with the corresponding customer experience departments and realised there were many processes that can be automated to help customers. Manish is also a serial entrepreneur who previously founded companies like Cellz and CorrZ in Noida.
With Rashi having a PhD in NLP and machine learning and Manish’s experience as a product manager and serial entrepreneur, the couple decided to take the startup flight. The couple bootstrapped the company with an initial investment of Rs 50 lakh. Today, Rezo.ai team has 10 employees.
Rezo’s training model operates with brands that already have a history of conversations ready, and also the ones that don’t have any customer support chat history.
Rashi says, "Suppose a brand wants to launch its WhatsApp channel and has no past conversational data, we help them generate the data. When the data is generated, we use it to adapt to our machine and train the model. Usually, our training takes three to four weeks for deployment."
The startup operates on a custom model where the pricing is case-dependent. The amount a brand has to pay depends on the volume of the conversations. Manish says it is a heavily discounted model without having to worry about the spike in the volume.
Speaking about how he went about on-boarding some initial customers, Manish says, "With all the financials agreed upon, we told them we won’t charge them until they are convinced with our pilot."
The duo also went through their inner networks and tried convincing people to use the platform. But what yielded better results was conducting free pilots with the customers.
Manish says, "It is typically based on how much a brand would have to spend if they were to do everything manually. We have factored in elements like lag-time (latency), 24x7 support, which requires a team, HR costs, and sitting costs. Putting everything together, we identified the most appropriate and cumulative cost that a brand would have to pay."
Personalisation is the key
The Rezo system is personalised in a way that it caters exactly to what the customer asks for without having to come back to a human for queries.
Rashi says, "A typical chatbot would ideally be keyword-based and it handles two or three or maybe five scenarios. Our system will generate nearly 1000’s of scenarios, thanks to Machine Learning. And hence, the quality is far better."
As for the customer experience, one of its customers, CarDekho, has a WhatsApp number, to which customers can send their messages or queries to. The Rezo.ai system, which sits behind this WhatsApp number, picks up the messages and understands the context of the query - what cars are they talking about, the preferences they have, etc.
If a customer is trying to evaluate between two cars, he/she can ask for pictures, videos, the price, and so on. Based on the query, the system will respond. The user can even book a test drive through the WhatsApp number.
Anurag Jain, Co-founder and COO at CarDekho said, "In the face of volatile change and rising customer demands, companies like us have an increased focus on maintaining and improving current levels of performance. It is more evident than ever that with evolving customer expectations, a superior experience that offers a real-time and informative response to queries becomes imperative."
Market scope and what’s ahead
The global market for conversational AI, which stands at $4.2 billion in 2019, is expected to grow to $15.7 billion by 2024, at a CAGR of 30.2 percent. Some of the prominent players in the Indian conversational AI space are Niki.ai, Haptik, Frontdesk.ai, Leena AI, Mihup, and Racetrack.
Manish says there are a dozen more pilots in progress, and he is hoping to convert them into full-fledged customers in the coming six to eight weeks.
The team claims to have clocked a revenue of $150,000 in the last financial year. Manish says they are also on the path to achieve an ARR of $1 million by the end of this financial year.
Manish is optimistic about the future, and his first plan is to partner with global resellers and partners.
"We identified a couple of them in the US. Further, we plan to expand pan-India for the next 12 months, as the opportunity is ample. And then, we want to hit the international market," he said.
(Edited by Megha Reddy)