In 2009, Rabi Gupta was watching a documentary about the Beijing Olympics when television suddenly seemed lonely. The program was reaching an enormous audience, yet Gupta could not talk with friends who were watching elsewhere. He imagined a social layer running alongside the broadcast. It was an ordinary viewer’s irritation, which is often where product stories hide before somebody gives them a company name.
Gupta was then an engineer at Wipro. He took the thought to his friend Ashish Kumar, and the pair began working from Gupta’s three-bedroom apartment in Noida. Family and friends supplied roughly ₹5 lakh to ₹7 lakh. The first product, iDubba, offered television listings, alerts and recommendations. It was useful, but utility was not the original spark. The spark was conversation.
So the product changed. In May 2013, iCouchApp arrived as a second screen for television. Viewers could chat while a show played, vote in reality programs and see messages appear as broadcast tickers. Zee Café became an early major client during a season of Grey’s Anatomy. What began as a guide turned into a live room, only the room was distributed across thousands of sofas.
The first cheque is a small piece of philosophy
The company’s first revenue took years. Gupta treated the cheque less like treasure than a laboratory result. “A real business as they say is incomplete without revenues,” he said at the time. However small, the payment showed that product development was facing in the right direction. The line is practical enough to print above a founder’s desk, preferably near the part where wishful thinking accumulates.
“One of the most important thing in a startup is to keep learning and growing. You need to be really flexible, optimistic and keep your eyes open for opportunities.”Rabi Gupta, during the iDubba years
Flexibility was already visible in the business model. iCouchApp served viewers, while the company sold broadcasters a mixture of product, engagement and analytics. Gupta said the revenue model emerged from a random meeting with Raghu Venkatraman. The lesson he drew was to keep meeting people, listen, and remain willing to alter the plan. A strategy can be rigorous without pretending it was delivered on stone tablets.
By the time Vidooly acquired iCouchApp in 2016, the platform had 1.5 million monthly active users. Gupta and Kumar had moved from a Noida apartment to relationships with major Indian television networks. The acquisition gave Gupta a second beginning. He moved to Silicon Valley, where the density of opportunity came with a personal inconvenience: he knew almost nobody.
The gifts were carried by hand
The next product began with the problem of making relationships in a new place. Gupta sent gifts to people he met, then discovered that choosing and delivering something personal was clumsy. EvaBot grew from that friction. The early team did something pleasingly unscalable: they personally delivered gifts. Gupta later recalled watching surprise on recipients’ faces. That reaction was the evidence. Software would need to remove the sender’s work without removing the recipient’s feeling.
EvaBot chatted with recipients to learn preferences and delivery details, then matched and fulfilled a gift on behalf of the sender. Gupta called the gift a “trojan horse” for learning useful information. The phrase sounds mischievous, but the product theory was direct. One thoughtfully phrased question can produce several pieces of context. A recipient gets something suited to them; a business learns how to communicate like it has been paying attention.
The team experimented in public. Install the Slack bot, receive an offer for a free coffee, start chatting. In the first three months, Gupta said EvaBot handled more than 500 gifts through organic word of mouth. In another early account, the company reported more than 3,000 gifts worth $80,000 over ten months. The platform later served corporate gifting programs, joined the Boost VC accelerator, and raised an $8.33 million Series A in 2022. Disclosed funding tracked for the company reached $11.13 million.
“I want to be able to give someone great gifts but it’s humanly impossible for people to do that.”Rabi Gupta on thoughtful relationships at scale
Gupta’s preferred description of AI was already less about replacing a person than furnishing one. He argued that machines could give people tools to be more thoughtful, especially as work became remote. His view of leadership followed a similar progression. At a small company, a founder does everything. As the group grows, the work becomes teaching others, then coaching leaders who can teach again. Scale, in this telling, is the art of giving away useful context.
From the gift to the reason for calling
EvaBot’s focus gradually widened beyond the gift. The company had learned to collect signals, personalize an interaction and help a seller build rapport. Revenoid is the current expression of that work: an AI system for enterprise sales teams. Its product language has changed from delight to pain, which sounds like a dreadful trade until you remember that sales begins with a customer problem.
Gupta argues that much sales software optimizes tactics: a list, a job-change alert, a line of personalization, a sequence. Tactics expire when buyers adapt. He wants Revenoid to work from what he calls a fundamental: identify strategic business pain at the right moment, then organize research and action around it. A hiring spike is a signal. A company restructuring its revenue operation is an initiative. The second has stakes; the first merely flashes.
The volume trap
Company-reported example: one rep booked three meetings from eleven researched calls. The bars compare unlike units and illustrate the contrast, not a common statistical scale.
One experiment sharpened the point. Gupta described an operation making more than 1,200 calls a day with a 1.3 percent connect-to-meeting rate. Then co-founder Satwick Saxena noticed a rep who had booked three meetings in the first half of a day after only eleven calls. The rep was selecting priority accounts and spending three or four minutes with Revenoid’s context before dialing. His opening demonstrated that he understood the prospect’s business, rather than merely announcing the interruption.
Gupta says the team became preoccupied with signal quality over call volume. It is a timely position because AI has made generic outreach cheap enough to exhaust everyone. When every seller can generate a competent paragraph instantly, another competent paragraph becomes wallpaper. The scarce item is the reason this particular buyer should speak with this particular seller now.
The quiet agent wins the morning
By February 2026, Gupta was writing about custom agents built around narrow customer workflows. One Revenoid agent scanned job boards for an enterprise IT staffing company, judged openings against that client’s ideal-customer logic and placed qualified leads with context inside the CRM. Gupta reported daily usage above 90 percent. The revealing detail was not the number. Reps forgot the agent existed. They arrived to find useful work waiting.
That kind of invisibility is a product compliment. Gupta’s first company placed conversation beside a television show. EvaBot placed preference gathering inside a gift exchange. Revenoid tries to place research inside a seller’s ordinary tools. Across the three, the technology behaves best when it stops demanding a ceremony of its own.
Gupta still writes like a founder running experiments rather than a commentator admiring abstractions. He asks readers to poke holes in his theories. He reports when a broad product gives way to a customer’s request for a precise workflow. He has also retained the optimism of his iDubba interviews, though it comes with the old cash-in-the-bank joke attached. Enthusiasm is useful; evidence pays the server bill.
His career offers a clean idea to steal. A founder’s durable asset may not be the product, market or pitch. It may be the question that keeps producing worthwhile products. Gupta’s question concerns attention: what does a person need to know, at this moment, to create a better interaction? For television viewers it was who else was watching. For a gift sender it was what the recipient liked. For a seller it is where the buyer’s business hurts.
Sixteen years after that Olympics documentary, the screens have changed and the audience has put on business attire. The gesture is recognizable. Somewhere, two people are about to speak. Gupta wants the machine to make the conversation less accidental.