Gaurav Singh Spent a Decade Teaching Machines to Answer the Phone

The useful thing about a customer asking the same question for the hundredth time is that, eventually, someone gets annoyed enough to build a company. In Bengaluru around 2015, Gaurav Singh was helping run the chat operation behind MagicTiger, a personal concierge service. The work had the glamour of a blinking cursor. People wanted orders, updates and answers; operators typed variations of the same replies. Repetition piled up like dishes after a banquet.

Singh saw software hiding inside the chore. The team began automating pieces of the conversation, first for itself and then for the broader market. That practical fix became Verloop.io, a customer-support automation company built before conversational AI became a compulsory phrase in every enterprise presentation. The category would change names, models and expectations several times. The underlying annoyance remained remarkably loyal.

By July 2026, Verloop was processing more than 20 million customer interactions each month. Nurix AI acquired the company to join its voice platform with Verloop's chat automation, multilingual experience and customer base across India and the Middle East. Singh joined Nurix's leadership team to work on product strategy, enterprise business and AI agents. An internal tool had taken the scenic route to becoming part of a larger enterprise platform.

A computer in Rewa

Long before enterprise software, there was Rewa, Madhya Pradesh. Singh grew up there and attended a local government school. He has recalled discovering coding around age ten and making projects by thirteen. In college at Jabalpur Engineering College, where he studied engineering from 2008 to 2012, programming also became a ticket out. He freelanced as a web developer to earn enough money to move to Bengaluru and begin a professional career.

His early jobs formed a compact tour of the web stack. At Tata Consultancy Services he worked on user interfaces and mobile web applications. At LT Research he tackled front-end performance, browser quirks and complex interfaces. At Vizury he built client-facing web applications. None of these roles reads like a dramatic origin myth, which is precisely why they matter. Verloop would later need an operator who understood that small technical choices become large operational consequences when customers arrive by the million.

The founder impulse appeared early. Singh has discussed an initial project called StockNinja, and in 2015 he co-founded GoDeliver, a hyperlocal service that let people request goods and services through chat. MagicTiger acquired it that July. Singh joined as a founding engineer. One startup exit led directly into the operation that exposed the next problem.

The first yes

Young software companies are often told to find product-market fit, a phrase that makes the process sound like locating the right socket. Singh's version was more human. Early customers became design partners. They showed the team where the product failed, what enterprises actually required and which requests were merely decorative.

Nykaa supplied the first consequential yes. Singh later remembered Sanjay Suri and Gaurav Tejwani trusting a tiny startup to build customer experience on chat before the choice looked obvious or safe. Verloop stayed with Nykaa through the retailer's path to an initial public offering. The relationship became a durable example of how an early customer can do more than purchase software: it can lend a young company enough reality to improve.

“Every founder remembers their first yes.”Gaurav Singh, reflecting on Nykaa

Reality was needed because the market initially treated conversational AI as a dressed-up FAQ page. Singh's team had to explain why a structured conversation, integrations, machine learning and a graceful handoff to a human agent added up to something more useful. They were building the category and the product in the same meeting. This is expensive education, particularly when the students control the budget.

At TechSparks in 2018, Singh turned his mistakes into a list for other first-time SaaS founders. Define the customer clearly. Resist decorative features. Do work that will not scale. Do not hire in a rush. Understand that revenue can arrive with very different costs. Above all, study customers more closely than competitors. The advice had the practical severity of someone who had already paid the tuition.

Three waves, one promise

Verloop's product changed as the technology did. Rules and workflows gave way to more capable natural-language systems. Machine learning improved intent recognition. Generative AI introduced flexible answers, document cognition, summarisation and tone controls. Voice joined chat. The company added support for more than 80 languages and 100 dialects, along with integrations, analytics and the quiet plumbing enterprises demand.

Singh has said the company survived three major technology waves and rebuilt its core through each of them. This is a more revealing achievement than merely adopting the newest model. Millions of daily conversations kept running while engineers altered the engine. For a consumer novelty, a wrong response can be funny. In a bank, retailer or telecom system, it can become a complaint, a compliance problem or a headline.

That explains Singh's insistence on human handoffs. He has argued that automated systems are well suited to routine questions, while people remain necessary for nuance, judgment and complex problem-solving. His preferred future was not a contact centre emptied for a press release. It was a system in which software handled repetition and routed uncertainty to the right person with context intact.

The philosophy can be stated plainly: solve the customer issue before admiring the tool. Verloop's later roadmap included quality control for call centres, easier knowledge transfer, agent assistance and reliable multilingual voice. The technology was interesting because the queue was still annoying.

The honest acquisition post

When the Nurix deal arrived in July 2026, Singh declined the polished victory note. His announcement remembered the years when chatbot was a punchline, the early believers, the difficult quarters and what he called more near-death experiences than he would publicly count. He thanked GrowthStory's Ganesh Krishnan and Srinivas Anumolu, later investors, and the people who carried the company through work outsiders never saw.

A second post was devoted almost entirely to colleagues. Singh credited CTO Ravi Petlur with rebuilding the engine in flight, Kumar Gaurav with turning messy requirements into simple products, Nikhil Gupta with retaining customers, Ankit Sarawagi with handling financial and operational fires, and Kiran Prabhu with selling enterprise AI when doors were firmly shut. The roll call said more about his operating style than a page of leadership adjectives could.

“The best part of this deal isn't the outcome. It's that this team stays together.”Singh after the Nurix acquisition

Nurix brought a complementary bet. Its NuPlay platform focused on voice AI; Verloop brought mature chat automation, multilingual capacity and customers in sectors where production reliability is part of the product. The combined company reported an annual run rate just above $10 million, with a plan to automate interactions across voice, chat and operational workflows. Financial terms of the acquisition were not disclosed.

Singh did not sell and retreat to a beach, at least not in the public version of events. He joined the Nurix leadership team to shape product strategy, enterprise go-to-market and the next generation of AI agents. His stated ambition is to help enterprises deploy autonomous systems across the customer lifecycle with the reliability that demos politely avoid.

The long game in tiny increments

In a 2025 long-form conversation, Singh connected his childhood in Rewa, early projects, college freelancing and founder years to an idea of focused, incremental progress. He spoke about building without waiting for validation and about the difference between startup glamour and startup life. The arc is easy to romanticise after an acquisition. It looked less tidy while invoices, hiring decisions, model changes and enterprise deployments arrived on schedule.

His public persona combines technical curiosity with a certain allergy to ceremony. He writes about obsession, but repeatedly returns to customers. He discusses artificial intelligence, then asks whether it works in production. He announces an exit, thanks the people who handled the 2 a.m. problems and closes with three words: “back to work.”

There is a useful lesson here for anyone building in a noisy category. Fashion can help a company get a meeting. Only operational depth keeps the customer after the meeting has forgotten why it was excited. Verloop began before conversational AI had its current glamour, persisted as the vocabulary changed, and found a buyer when chat, voice and workflows began converging into one enterprise surface.

The story starts with a repeated question in a chat window. It ends, for now, with a larger company trying to make millions of conversations reliable across channels. Somewhere between those points sits Singh's real product: the habit of noticing boring work, staying with it and rebuilding whenever the machine learns a new trick.