Verloop.io Turned 100 Chat Agents Into a Bot, Then Sold the Bot

The best product Verloop.io ever built was not the one it set out to build. Around 2015, a small team in Bengaluru was running MagicTiger, a personal-concierge app. To keep up with demand, more than a hundred people sat in a chat operations room answering the same customer questions over and over. It was expensive, repetitive, and, if you were paying attention, an obvious problem to solve with software.

So the team started automating its own desk. First the boring replies, then the routing, then the follow-ups. And somewhere in that grind they noticed something more interesting than the concierge business itself: the automation was the company. In July 2026, that realization closed a loop that started a decade earlier, when Nurix AI acquired Verloop.io to bolt proven voice and chat automation onto its enterprise platform.

01A pivot most founders only talk about

Every founder claims they will pivot when the data says so. Very few actually do it, because pivoting means admitting the thing you pitched is not the thing that works. Gaurav Singh and Piyush Mishra had a B2C app that, like most B2C apps, was hard to make profitable. What they also had was a working piece of infrastructure they built for themselves. The unglamorous decision was to close the front door and reopen it as a B2B one.

The name gives away the idea. Verloop is a play on "verbal loop" - the back-and-forth of a conversation. If you could handle the loop reliably, at scale, in a customer's own language, you had a product that every consumer brand needed and almost none wanted to build in-house.

Singh is not a typical demo-stage founder. He fell in love with coding at 13 and spent his career on scalable systems and personalization - the plumbing, not the pitch. That temperament shows up in the company. Verloop.io never tried to be the loudest voice in a very loud category. It picked one promise and kept it: make the repetitive 80% of support tickets resolve themselves so human agents only touch the conversations that actually need a human.

02The unsexy moat

Customer support is the least glamorous department in most companies. Nobody puts "we answered 40,000 password-reset questions" on a keynote slide. But that is exactly where the money was, and Verloop.io built its business on the part of the work everyone else found boring.

The platform grew to cover voice, chat, and messaging in more than 80 languages, sitting quietly behind the support windows of brands you have probably used - Decathlon, Adani Capital, DSP Mutual Fund, Scripbox, Rentomojo. When infrastructure works, it disappears. That invisibility is precisely what made the company acquirable: it was already load-bearing for thousands of brands.

In 2024 the company shipped a tell of where its head was at: Sparks, a generative-AI tool that auto-scores support conversations for quality. It is a subtle move. Automating a reply is the easy half. Knowing whether the automated reply was any good - polite, correct, on-brand - is the hard half, and it is the half that keeps enterprise buyers up at night. Building the grader, not just the answerer, is a sign of a company that had actually run a support desk.

03A crowded, loud market

By the mid-2020s, conversational and voice AI was one of the most contested corners of software. Verloop.io was up against names with far bigger budgets: Google with Dialogflow and its contact-center stack, Five9 baking virtual agents into its call-center platform, PolyAI chasing uncannily natural phone calls, and builder platforms like Kore.ai and Botpress. Add regional rivals - yellow.ai, Ada, Haptik - and the noise was constant.

Verloop.io's answer to the noise was to not compete on noise. It did not try to out-feature Google or out-spend Five9. It stayed focused on the promise and let the badges accumulate - a Leader placement in the G2 Spring 2025 report, 50+ badges across categories from conversational support to live chat. Modest funding, roughly $15.3M across six rounds from 31 investors including Falcon Edge Capital, IDFC Parampara, Alpha Wave, Kris Gopalakrishnan, and K Ganesh. No unicorn headline. Just a company that knew what it was for.

04The exit, and what it says

In July 2026, Mukesh Bansal's Nurix AI acquired Verloop.io. The logic is clean: Nurix wanted enterprise-grade voice and chat automation, and Verloop.io had ten years of it already running in production. Singh did not cash out and vanish - he joined Nurix AI's leadership to work on product strategy, enterprise expansion, and the next wave of AI agents. The operator came with the software, which is usually a sign the buyer wanted the judgment, not just the codebase.

There is a lesson in here worth stealing, and it has nothing to do with AI. Your worst operational headache is often somebody else's product. The workarounds you build to survive - the scripts, the macros, the internal tools nobody outside your team ever sees - are the raw material of a real business. Verloop.io is a ten-year argument for paying attention to your own grunt work.

The tidy arc reads like this: born from a pivot, funded modestly, focused ruthlessly on the boring parts of support, acquired by a bigger player who wanted exactly that. No drama, no hype cycle chased. In a category built on demos, Verloop.io built on tickets - and the tickets turned out to be the moat all along.

Q&ACommon questions