Experiential commerce is what comes after quick commerce

The Milli Vanilli effect and the collapse of online trust

Muzamil opens the episode with his standing frame: AI is inside a financial bubble, but the underlying technology is durable. What caught his attention about Epic was not another storefront but a reversal, a company using AI to shrink the catalogue and then sending a human to the customer’s door within an hour to actually demo the product. The numbers are the reason the reversal is worth taking seriously. Epic launched about eight months before the recording, raised a million dollars in pre-seed, and has since reached $4 million ARR, with a target of $30 million by the end of the year, all from two cities: Bangalore and Delhi NCR.

Varun picks up the thread with a borrowed line from Mark Cuban about a “Milli vanil effect” in the post-AI era. Once anyone can spin up an Apple-adjacent brand, a Shopify store, AI avatars, and paid reviews within minutes, the internet becomes untrustworthy by default. “As internet gets lower on trust because of commoditization of content through AI,” Varun argues, society is already voting with its feet toward concerts, dining, and community. Epic’s bet is that commerce follows the same gravity.

He lays out the commerce timeline the way he sees it: brick and mortar in the 90s and 2000s, Amazon-style e-commerce in the 2010s, quick commerce after Covid turning into “multi-billion dollars in four to five years.” What comes next, he says, is experiential commerce, specifically for “high touch, high consideration, high AOV products.”

What actually happens in 60 minutes

The mechanic is deliberately simple. A user lands on Epic, talks to an AI agent by voice or text, and shares a use case: a robo vacuum for a specific problem, a coffee machine within a budget, a phone between two brands. The agent narrows the internet down to two or three products. A single button, free home demo in 60 minutes, then triggers physical fulfilment: an Epic operator picks the product up from a nearby store, arrives at the customer’s home within the hour, and runs a real demo.

If it is a coffee machine, they make coffee. For a guitar, Varun says, “we’ll bring two guitars, a person will sing three songs for you and then you can actually buy them.”

The category choice is economic, not aesthetic. Epic started with electronics because “anything above a $100 to $150 USD is where the economy start making sense.” Below that threshold, Epic routes to virtual demos, closer to the Chinese live-commerce model, because sending a human into a living room does not pencil.

The AI stack behind the storefront

Muzamil pushes past the customer-facing layer to ask what AI is doing in the engine. Varun’s answer is unusually concrete. Epic runs sales AI agents on the site, tele-calling AI agents, internal apps stitching logistics to demo completion, and, critically, consented voice recording during the in-home demo itself. That recording becomes primary data: what the consumer actually cares about, which objections come up, what the next pitch should be.

The build side is equally lean. “Our tech team is very lean like five member team and they all are on cloud code,” Varun says, describing engineers who solve operational and GTM inefficiencies as they surface. He mentions discovering tools like Noodleseed on Product Hunt, and Gojiberry, among the automations they scan for internal use.

He then extends the arc. Referencing a 314 Capital presentation on labour disruption, white collar first, blue collar next as humanoids mature, he floats a two to three year horizon in which Epic itself might pilot humanoid demo operators or offer customers a choice between a human and an AI demo. Not inevitable, he says, but a natural extension of what Epic is already doing.

A counter thesis on Google versus LLMs

Muzamil raises the shift directly. India is now among the largest consumer bases for LLMs in the world, with tens of millions of daily chatbot users, and discovery is visibly moving off Google. He asks whether Epic has begun optimising for AEO and GEO.

Varun’s response is a genuine contrarian frame. “While it looks at the surface that Google is being disrupted,” he says, “I have a counter thesis.” His argument: LLMs are still scraping an index Google shaped, and once the current funding cycles stabilise, “Google will still have a lot of power in terms of even if you are searching through LLMs what they are eventually showing you are some most relevant links.”

Where he does concede full disruption is at the consumer interface. Ninety-five percent of searches, he estimates, are moving toward chat-native experiences. The important shift is not who owns the index, it is that discovery is becoming phrase-led rather than keyword-led: “how we converse is how these engines are getting built.” Epic’s response has been to plant its differentiator, free home demo in 60 minutes, into content structured around the way people actually ask LLMs for buying help.

Unit economics before scale

Muzamil pushes on the part of the story that usually breaks in Indian consumer tech: the bleeding. Global-south markets are price-sensitive, quick commerce left a sour taste post-Covid, and the pattern is usually a market-share war funded by investor capital. He asks the direct question: is Epic profitable per order, or is this a subsidy game waiting for a Flipkart to squash it?

Varun’s answer separates Epic from the reflex. “The bootstrapped company and those values principles that I’ve learned,” he says, referencing the exit from his previous D2C-focused agency, shaped the discipline. Epic is CM2 positive per order today, running roughly a 20 to 25% commission model, and has not yet introduced a convenience fee even though the service, a person arriving at the door in an hour, clearly justifies one.

The more interesting layer is what Varun calls additional monetisation surfaces once demo volume compounds. Every operator entering a home is a channel. The T-shirt is sponsorable, the bag is sponsorable. When Apple launches a new phone, Epic’s demo fleet could carry a demo unit into tens of thousands of homes a day, targeted against known household profiles by pin-code. And the consented voice data from demos is, on its own, a defensible asset. “Scale is the oxygen for a marketplace,” Varun says, quoting his investors at Infoedge.

India as the application layer

The conversation broadens. Muzamil, who has lived across Pakistan, Dubai, and the US, tells Varun that Asia now feels more like the land of opportunity than his imagined America, vibrant and problem-rich. He asks Varun to describe the Indian startup ecosystem honestly, past the polarised narratives of “rising India” and “valuation bust.”

Varun’s read is that the money loop has closed. The Flipkart-Walmart exit “made hundreds hundreds if not thousands of people millionaires,” and those operators are now backing founders. Government policy, taxes, and subsidies are aligned. Shark Tank India shifted the cultural aspiration, in his framing, from becoming a Sachin Tendulkar to becoming an entrepreneur, reaching tier-2 and tier-3 cities.

India is not building the silicon layer, he concedes, but “the application layer is beautifully getting built,” and half of San Francisco, he notes, now has a CTO or engineering team in Bangalore. His comparison is direct: “It’s like what happened with China 10 years back.”

He closes the personal loop honestly. He spent a year in Germany during his master’s, loved the balance of European life, but chose India because the pace matched his stage. He still plans to end up with “a few tennis courts” in Italy at 50. But not yet.

The telescope and the microscope

Muzamil asks about the next five years. Two cities today, Bangalore and Delhi NCR. Where does this go?

Varun outlines a tiered expansion inside India first, top five then top ten tier-one and tier-two cities, because “the top 10% of the economy actually or of the population holds a lot of money in their hands.” Internationally, the model is being pulled. Friends in Dubai and Singapore describe the same demand pattern: families with kids, older customers, high-consideration purchases. Epic’s current round even includes a Japanese investor opening a corridor to bring Japanese brands into India through home demos, on the thesis that “Japanese technology cannot be explained through an online ad.”

The framing he uses for how founders navigate this is worth pulling out directly: “one hand we hold a telescope, one hand we hold a microscope.” The telescope says experiential commerce is a ten-year wave. The microscope is who to hire next, which product to launch next. Both have to be held at the same time.

Startups Varun is watching

Toward the end, Muzamil asks Varun to name Indian startups that are not yet famous but that he thinks matter. Varun picks four across categories. Superu, a protein bar built around a yeast-derived protein that avoids bloating, distributed through quick commerce. Rapido, which found a business model that took share from Uber and Ola after five to six years of head start. Server AI, building Indian AI infrastructure and models. And a broader gesture at Indian space-tech and defence-tech startups, which he notes take unusual guts to start in an economy where capital for hardware and deep tech is not “exponentially available.”

Muzamil connects it to a recent study he did on SpaceX, noting how many space startups either come directly from India or have Indian co-founders, framing it as the same pattern that made Indian IT world-famous, now repeating in frontier categories. By the end, the argument Varun has built is clean: AI hollows out trust online, experience becomes the scarce good, and whoever owns the last-mile experience layer, and the data that comes from being physically inside the customer’s home, owns the next decade of high-consideration commerce.