Lumio’s Project Neo hands-on: How WhatsApp and Instagram are becoming your TV remote

The race to bring AI into the living room is heating up. Amazon is preparing to bring Alexa+ to India, Google is embedding Gemini across more devices, and every major tech company is trying to figure out how AI will change the way we interact with our televisions.

At the same time, agentic AI is emerging as one of the biggest technology trends in India. We've already seen its early applications in e-commerce, where companies like Amazon, Zepto and Swiggy are building AI agents that don't just answer questions but actively complete tasks on behalf of users. We've covered that shift in detail separately. Check it out here.

Lumio believes the next frontier for agentic AI isn't shopping. It's television.


During our visit to the company's Bengaluru office, we sat down with Raghu Reddy, CEO of Circuit House Technologies, and Kailash Sankaranarayanan, Co-Founder & COO, to understand the thinking behind Project Neo, Lumio's new AI-powered content discovery experience. After trying the public beta, it's clear the company isn't trying to reinvent the TV. It's trying to reinvent how we discover what to watch.

Your next TV remote might be WhatsApp

Project Neo is built into Lumio's TLDR platform and works through WhatsApp and Instagram. Instead of picking up your remote, typing with a D-pad or jumping between streaming apps, users can simply message Neo in natural language.

"Show me a thriller."
"I'm bored."
Or even share an Instagram Reel, movie poster or screenshot.

Neo understands the request and sends the content directly to a Lumio Vision TV or Arc projector through the TLDR app.

It's a simple idea, but one that feels surprisingly intuitive.

Why Lumio isn't worried about Alexa+

One of the first questions we put to the team was about the growing competition. With Amazon's Alexa+ expected to arrive in India and Google continuing to expand Gemini across Android and Google TV, where does Project Neo fit?

Rather than trying to compete with general-purpose AI assistants, Lumio believes it's solving a much more specific problem: helping users move from discovering content to actually watching it.

"The goal when we started TLDR was simple: how do you reduce the time it takes for someone to find what they want to watch? We think most TV operating systems have different objectives, whereas ours is focused on getting customers to the content they actually want as quickly as possible."

Raghu Reddy, CEO, Circuit House Technologies

The company's view is that the biggest pain point isn't controlling the television. It's remembering what you wanted to watch.

"Discovery happens everywhere. A friend recommends a show, someone shares a trailer, or you're scrolling Instagram. But when you finally sit in front of the TV, you've forgotten all about it and end up scrolling. We're trying to bridge that gap by bringing the agent to where people already are, on WhatsApp and Instagram."
Kailash Sankaranarayanan, Co-Founder & COO, Circuit House Technologies

More conversation, less scrolling

Unlike traditional TV search, Neo doesn't expect users to remember exact titles.
Users can search by actor, director, language, genre, mood or even describe a half-remembered plot. If someone simply says they're bored, Neo recommends trending content across supported streaming platforms.
Instagram integration goes a step further. Share a movie poster, trailer or screenshot with Neo, and it identifies the title before sending it directly to the TV.

"It's not the typical search behaviour. It's conversational. You can type naturally, use colloquial language or simply forward something you've seen. The agent figures out what you're referring to and brings it to your TV."
Kailash Sankaranarayanan

What's happening behind the scenes?

One of the questions we were particularly curious about was whether Neo was built entirely in-house or simply wrapped around an existing large language model.
Lumio says it's using capable AI models but has built its own intelligence layer specifically for entertainment discovery.

"We're not developing foundation models from scratch. We use capable AI models, but there's a lot of work in fine-tuning them for entertainment discovery. On top of that, we've built our own tools that understand streaming catalogues and help surface the right content."
Kailash Sankaranarayanan

The ambition, however, goes beyond simply answering search queries.
"Over time, Neo should become more like a tastemaker. There's a human element to curation that people still value, and we're trying to build that intelligence into how the agent evolves."
Kailash Sankaranarayanan

Cloud AI, not heavier TVs

Another question we raised was whether bringing AI to televisions would require significantly more powerful hardware, especially at a time when rising memory prices are affecting consumer electronics across the industry.
Raghu Reddy acknowledged that the entire industry is feeling the pressure of higher component costs, but says Neo itself isn't dependent on more powerful local hardware.

"Memory costs have gone up across consumer electronics and everyone will feel that impact. But Neo itself runs in the cloud, so we're not relying on the TV's computing power for the AI experience."
Raghu Reddy

Privacy was another obvious concern.
According to Lumio, user identities aren't directly accessible even within the company.
"All user information is encrypted and hashed. We can understand the queries coming in, but they aren't identifiable to a specific user. Even internally, that's by design."
Kailash Sankaranarayanan

Software is becoming the differentiator

Perhaps the biggest takeaway from our conversation wasn't the AI itself.
It was Lumio's belief that while hardware may sell a customer their first television, software is what keeps them coming back. It's a strategy that mirrors the smartphone industry, where users often remain loyal to software experiences long after hardware specifications have become comparable.

"People initially buy hardware, but over time they become attached to software experiences. That's exactly what we're trying to build with TLDR and Neo. We want users to miss these features if they ever switch to another television."
Raghu Reddy

Alongside Project Neo's public beta launch, Lumio also revealed it has crossed ₹100 crore in gross merchandise value (GMV) just over a year after entering the television market, with nearly half of its customers switching from established TV brands.

Our take

Project Neo is still in beta, and like every AI product, it has room to evolve. But after spending time with the team and trying the experience firsthand, it feels like one of the more practical implementations of agentic AI we've seen in consumer electronics.

The television industry has spent years competing on brighter displays, higher refresh rates and faster processors. Lumio is betting the next battle won't be fought on hardware alone, but on reducing the friction between discovering something worth watching and actually pressing play.

Whether that bet pays off remains to be seen. But if agentic AI is truly about helping users complete tasks instead of simply answering questions, Lumio's approach feels like one of the clearest examples of that shift. The future of TV may not start with the remote. It might start with a WhatsApp message.