D2C brands take AI beyond the marketing function

AI is finding its way into the core decisions of D2C businesses, from what products brands build to how they understand customers and generate revenue.

At the e4m D2C Summit 2026, a panel featuring Vedansh Kumar, CMO, boAt; Midhula Devabhaktuni, Co-Founder & CMO, Mivi; Nikhil Sareen, Co-Founder, Qubo; Shuchi Pandya, Principal, Fireside Ventures; and Kanishk Arya, COO, SleepyCat, explored how AI is being deployed across the D2C ecosystem. Moderated by Ayush Bansal, Director, Liberty Shoes, the discussion moved from marketplaces and customer intelligence to AI-powered products, subscriptions, first-party data, acquisition costs and the limits of automation.

Kumar began by stating that a lot of the AI opportunity lies before the customer reaches the marketplace. “Today, our customer is looking at maybe six to 12 different digital services before arriving at what they want to buy, even on a marketplace. That’s where AI is helping us,” he said.

Hence, understanding what consumers are saying across those touchpoints and ensuring the brand is part of that consideration is challenging but crucial. Kumar pointed to queries such as the best headphones to buy under a certain price point and whether boAt is being cited across platforms such as YouTube or LLMs. “We want to be part of that conversation, understand what they’re saying and at what price point. That’s where AI is helping us win that battle.”

At boAt, AI is also being used to understand customers faster. “There are certain tools that we’ve built and are working with which help us understand long-term brand equity scores, which used to be unlocked only by traditional brand tracks. Now, that’s being done at scale, much faster, and we’re synthesising that data into very actionable items and line items that we as a brand should be in service of,” Kumar said.

Another area of application is testing. Instead of relying only on focus groups, brands can now use simulation environments to test products, CMF (Colour, Material, and Finish) designs, pricing features and content before they reach the market.

For Mivi, AI has become a part of the product itself. Devabhaktuni highlighted the company’s AI earbuds and Mivi AI, describing the approach as a way to change how consumers perceive the brand. “Being able to take AI and integrate it into the product and to be able to move at a fast pace has enabled us not just to create a platform and offer the consumer a different product, but also to create a different perception of the brand,” she said.

The company has taken that approach across its operations as well. “Every department has moved to AI. We’ve built an end-to-end custom ERP system within four months using AI,” Devabhaktuni said, adding that Mivi also has dedicated AI teams. She can see the results in a combination of stronger customer understanding and the ability to deliver faster.

Next in the discussion, Sareen shared Qubo’s initiative to use AI to turn hardware into an ongoing relationship. Qubo’s devices are designed to actively respond to threats rather than simply record them. If an intruder enters a home, the device can trigger an alarm, switch on its spotlight and call the customer.

Edge AI built into the device has created a subscription opportunity. “We’re able to build this as a subscription feature behind the paywall and give customers enough value that they can actually pay for this on a monthly or annual basis,” Sareen said. He sees the model altering the economics of customer acquisition. Recurring revenue can help subsidise the initial hardware purchase or justify spending more to acquire a customer whose value extends over a longer period.

From an investor’s perspective, whether AI produces a meaningful business benefit is what matters most. Pandya grouped its strongest applications into three areas, namely saving time and creating convenience, gathering knowledge and customer insights, and improving decision-making.

She cited a portfolio company that used AI across its recruitment process, taking 1,000 to 2,000 applicants for a founder’s office role down to around 15 candidates before the interview round. “AI had done a brilliant job of getting that 1,000 down to 15,” she said.

But with AI becoming ubiquitous, Pandya cautioned founders against treating it as an answer in itself. “AI is a very small word for something which is very large. People just throw it around a lot. Intelligent uses can actually make sense.”

Arya took the session forward by highlighting SleepyCat’s specific problem. They want to reduce the mattress purchase journey that can stretch to three weeks. He said the central issue is conviction, whether customers believe they are choosing the right product for themselves, whether it will function as promised and whether it will last. So, they are using AI to provide recommendations both on and off the platform, based on what customers are looking for, what they have slept on before and the problem they are trying to solve.

At boAt, AI is helping narrow the enormous number of choices involved in product development. Simulation tools are helping determine product colours, pricing and marketing content, while data from app and website usage is helping identify potential problems. “Half the decision is already made once you’ve decided what product you’re building. So, if you have an AI intervention there that can help you narrow down the choices, that’s a very valuable insight to have,” Kumar said.

Mivi has been more selective about where AI fits into strategic decisions. Devabhaktuni said the company still relies on insights from its e-commerce partnerships when deciding which categories to enter. Its focus is instead on using AI to make hardware more accessible. With Mivi preparing to launch its first phone, the company is exploring ways to build AI capabilities into the device.

“For us, rather than extensively using AI just to understand consumer behaviour, we look at ourselves as a hardware company. How can we take the software capabilities of AI and figure out how it can provide better insights to the consumer?”

Talking about the issue of privacy, Devabhaktuni said Mivi has deliberately limited the information it accesses, relying on higher-level insights rather than deeper individual data. “We will not tap into the consumer’s insights other than at a higher level,” she said, arguing that stronger regulation around consumer data will be necessary.

Next, the panel discussed the differences between building AI capabilities internally and buying them from outside providers.

Sareen discussed, “You should definitely not build everything. You should shop where it’s available. But deciding which between the two is critical.”

Qubo builds anything that directly affects the consumer experience internally, including firmware, the app and AI running on its devices. Generic foundation models and cloud services are explored off the shelf. The decision ultimately depends on the company’s stage, priorities and the role technology plays in its ecosystem.

That same principle applies to measuring AI’s impact. Arya pointed to customer value as the metric that brings conversion, CAC and other measures together. “The efficiency point comes into the picture when I can see that overall I’m able to get a greater share of wallet from my customer for whatever I’m spending,” he said.

Across the businesses represented on the panel, AI has already changed internal operations. Devabhaktuni said Mivi had reduced its team by around 40% while continuing to grow and prepare for its mobile launch. Kumar pointed to lower costs and faster execution across teams, while Sareen highlighted the ability to analyse thousands of customer conversations.

“Earlier, you would typically have a sales manager randomly listen to 10 calls in a day or 100 calls a week and then take insights from there. But now you’re able to listen to every single call, get a transcript, get insights, and pass those insights on to every different function within the organisation,” Sareen noted.

Yet the panel’s final message was less about what AI can automate and more about what it cannot replace.

Pandya warned that the biggest risk is losing the human element behind the technology. “The people who are consuming brands are still human. That’s not artificial. For brands to not lose that authenticity, it is important to still remain and have that empathy for the customer,” she said.

Sareen cautioned founders against being seduced by polished demonstrations of AI tools that do not always perform as promised once deployed.

Arya brought the discussion back to the role of brands themselves. “As brands, we still need to own the consumer understanding as well as have our own thought into the creative sharpness, which I think AI does an average job at,” he concluded. “Brands are built when somebody goes really deep and really sharp to solve a problem.”