Underneat Built a Superhero Campaign. The Superhero Was Our AI.

There's a cupboard somewhere in India where the clothes have opinions.

A backless dress named Kavya is in tears, because she's spent years in the closet watching saris and kurtas leave for Bandra weddings and Bali honeymoons while she stays behind. She has dreams. Bangkok. Bali. Even just Bandra. None of them happen. She's so done that she's seriously considering signing up to become a pocha.

Then a hat starts narrating. Then Kusha Kapila descends from the top drawer wearing brown leather and a cape. She points at Kavya, holds up a backless bodysuit, and the universe rearranges itself. The backless dress finally goes out.

This is the first film in Underwoman, the brand campaign Underneat launched on February 6, 2026. Conceptualized by One Hand Clap, written by Manaswi Mohata, directed by Devika Chaturvedi at Crazy Few Films. A multi-part cinematic universe rolling across digital, social, Nykaa, Myntra, Amazon, and Blinkit. Two more films are coming.

The campaign's core line is what Underwoman whispers to Kavya before she saves the day:

"I'll tell you what to wear UNDERNEAT."

That tagline isn't copywriting. It's a product behavior.

The Underwoman on the site is Manifest AI

Go to underneat.in right now. In the corner of the page, you'll see a small widget. It doesn't say "Chat with support." It doesn't say "Help." It says:

Chat with Underwoman.

Click it. You'll meet the same character from the film. She'll answer what fabric works in 38 degree heat. Whether the high-waist will roll down under a saree. Whether a particular bodysuit will hold under a backless dress at a sangeet. Compression level, breathability, butt-lift, side-coverage, daily wear viability. All of it, in seconds.

That's Manifest AI.

The character in the campaign and the product on the website are the same thing. Underneat didn't deploy an assistant and then build a campaign around it. They didn't build a campaign and then bolt on a chat widget to match. They built a brand identity around what the AI on their site was already doing for buyers every day. The film made literal what the product was already saying.

This is, as far as we can tell, the first time a D2C brand has built a full cinematic campaign with celebrity talent around the behavior of a SaaS product running on their store.

What follows is the case study. What the AI does, why it works, and what it delivered before the cameras ever rolled.

Who is Underneat?

Underneat is a body-inclusive D2C shapewear and innerwear brand built for Indian bodies, climates, and silhouettes. Co-founded by Kusha Kapila and Vimarsh Razdan, the brand runs a 15 to 20 day drop cycle across bras, bodysuits, shapewear, and essentials. Most discovery happens on Instagram, on the back of Kusha's creator audience. The category is high-anxiety and high-return: a buyer who isn't sure about size or fabric either bounces, or buys three sizes and sends two back. Confidence drives every conversion.

The Challenge

Shapewear isn't a comparison-shop category. It's a confidence category. Buyers want a fitting room consultation, not a checkout button. Underneat's existing setup wasn't built for that level of guidance.

• No AI on the website to consult on body type, compression level, fabric breathability, or use case
• WhatsApp flows were strong on transactional queries (orders, tracking) but couldn't advise on fit
• Instagram DMs were running on template replies that broke the moment a buyer asked something specific
• The support team was answering the same fabric, sizing, and coverage queries manually, all day
• No intelligent discovery flow to match a hesitating shopper to the right SKU before she left

What Underneat Deployed

Underneat layered Manifest AI on top of Bik's existing infrastructure. Three things changed across the buyer journey.

The product page is where shapewear sales live or die. Manifest AI deployed a product assistant directly on every PDP, capable of answering specific questions about fabric stretch, compression level, coverage, breathability, and use case. A buyer asking "is this bodysuit suitable for daily wear" gets a precise answer back, with context on fabric, support, and skin sensitivity. Floating nudges proactively surface the questions buyers are already hesitating over, so the AI starts the conversation before the back button gets clicked. Over the study period, the PDP layer drove 4,141 AI-led interactions across product pages.

Underneat's strongest discovery channels are also where most buyer hesitation lives. The Bik and Manifest stack upgraded WhatsApp from transactional to consultative. A buyer who comments "Price please?" on Instagram gets a real product recommendation in DMs, with pricing, an IG-only discount code, and a direct purchase link. The same intelligence runs through WhatsApp, where buyers can browse the catalog, get fit advice, and track orders inside one continuous thread. Across both channels, the system handled 5,071 conversations end-to-end, with 98%+ satisfaction.

Not every buyer converts on the first visit, and shapewear especially needs a second touchpoint. Manifest AI's lead capture flows convert browsing sessions into retargetable WhatsApp shoppers inside the conversation itself, no separate popup, no friction. A buyer asking "I usually wear L in jockey but it's loose, what should I get here?" gets a specific size recommendation, a product link, and a prompt to receive their fit guide on WhatsApp. The result: a retargeting pool of buyers who already know what they want, feeding the marketing engine directly. Shoppers who engaged with the AI converted at 8.4%, delivering a 22.3% uplift in assisted sales.

The Results (April to May 2026)

Revenue: The deployment drove ₹37L+ in revenue through Manifest and Bik, contributing 9.5% of total digital revenue with a 4.67X conversion uplift versus non-interactors.

Orders: Over 1,500 orders were driven by AI conversations, with chat interactors converting at 8.4% and AI-assisted sales seeing a 22.3% uplift.

Support: The team automated 5,200+ queries instantly, dropped repetitive customer queries by 60%+, and held a 98%+ satisfaction score across all AI sessions.

Engagement: 4,141 AI-led PDP interactions plus 5,071 end-to-end conversations across WhatsApp and Instagram, with 83% of sessions generating active replies and engagement.

Why this matters

Most SaaS products get deployed. The good ones get adopted. A rare few get integrated so deeply into how a brand actually operates that the brand starts building creative around them.

Underneat's Underwoman campaign isn't a metaphor for what Manifest AI does. It's a depiction of it. The character in the film tells women what to wear. The AI on the site tells women what to wear. The widget literally says "Chat with Underwoman." The tagline that closes the film is the same tagline that opens every product conversation on the website.

When Kusha Kapila and Vimarsh Razdan describe Underwoman, they describe Manifest AI:

"Underwoman is about those silent moments when you just want your clothes to work with you, not against you. Underneat is built on the belief that when you feel comfortable underneath, you don't need fixing. You just need freedom."

The numbers above are what that integration delivered before the campaign even launched. The 4.67X conversion lift, the ₹37L+, the 60% drop in repetitive queries. The campaign is what happens when a brand looks at its own AI's behavior and decides it's worth putting in a film.

Want this for your brand?

You won't need a creative agency or a celebrity to start. You'll need a product page where a buyer can ask a real question and get a real answer, in seconds, in the channel she's already on. That's where it begins. Everything Underneat did, including the cape, came after.

Sources and credits:

The Underwoman campaign was conceptualized by One Hand Clap, written by Manaswi Mohata, and directed by Devika Chaturvedi of Crazy Few Films.
The first film, "Kavya's Backless Dress," is live on YouTube and rolling across Nykaa, Myntra, Amazon, and Blinkit.
Campaign coverage: Adgully, Afaqs, MediaNews4U. Performance data sourced from Underneat's Manifest AI and Bik dashboards, April to May 2026.