How Havells Is Turning Your Home Into a Multilingual AI Assistant
Havells has rolled out a conversational AI voice feature in its mobile app that understands natural speech across eight languages, including English and seven Indian languages, allowing users to control connected home devices without memorizing commands. The feature, which went live for all 2.67 million Havells One app users, represents a significant shift from rigid, command-based voice interfaces to AI agents that interpret intent and handle code-mixed speech, where speakers blend multiple languages in a single sentence.
What Makes This Different From Traditional Voice Commands?
The previous Havells voice experience required users to remember specific phrases and worked only in English and Hindi on selected appliances. The new Agentic AI Voice feature, built using ElevenLabs' streaming speech-to-text and text-to-speech models, interprets what users actually mean rather than matching fixed syntax patterns. This matters because households in India often don't speak a single language; family members naturally code-mix, moving between three languages in one sentence. Instead of forcing users to learn how to talk to their devices, the technology now adapts to how people already speak.
The system can handle complex, multi-device requests. When a user says "make the room comfortable," the agent orchestrates the air conditioner, air purifier, fan, and lights together, executing commands across different connectivity protocols simultaneously. The agent also retains context across multiple turns of conversation, asks for clarification when commands are ambiguous, and can explain device error codes in plain spoken language.
Which Languages and Devices Are Supported?
The feature currently supports eight languages across a broad range of connected devices. Users can interact in Hindi, Tamil, Telugu, Bengali, Kannada, Gujarati, and Malayalam alongside English. The supported device ecosystem includes fans, air purifiers, water purifiers, water heaters, Lloyd air conditioners, washing machines, Crabtree smart switches, and upcoming smart home products.
Havells has signaled that the rollout is just the beginning. The company plans to expand support to additional Indian languages in the future and deepen personalization for individual household members, bring more appliances onto the voice platform, and extend voice capabilities beyond device control into product FAQs, troubleshooting, and scheduling.
How Did Havells Build This So Quickly?
The Havells Center for Research and Innovation (CRI) moved the solution into production in approximately six weeks, with support from ElevenLabs' Forward Deployed Engineering team on voice quality and regional language pronunciation tuning. The team evaluated voice AI models on recognition accuracy, multilingual support, latency, and voice synthesis quality before selecting ElevenLabs' technology as the foundation.
"People have always had to learn the interface of communicating: the commands, the syntax, the order they come in. That is hardest in India, where a household does not speak one language and often moves between three in a single sentence. Powered by ElevenLabs, Havells now runs seven Indian languages alongside English across an appliance range, so the person no longer has to meet the technology halfway. They say what they want to the agent, in the language they were already thinking in, and the devices respond," said Karthik Rajaram.
Karthik Rajaram, General Manager and Country Head, India, ElevenLabs
Steps to Use Havells' New AI Voice Feature
- Register Your Device: Ensure your Havells smart device is registered in the Havells One mobile app to access the Agentic AI Voice feature.
- Speak Naturally in Your Preferred Language: Use any of the eight supported languages, including code-mixed speech, without worrying about specific command syntax or phrases.
- Make Multi-Device Requests: Issue complex commands that coordinate multiple appliances at once, such as asking for comfort settings that adjust temperature, air quality, and lighting simultaneously.
- Ask for Clarification or Explanations: If the agent needs more information, it will ask follow-up questions, and it can explain device error codes in plain spoken language rather than cryptic codes.
Why Does This Matter for the Smart Home Industry?
This launch signals a broader shift in how consumers interact with connected devices. Voice has long been positioned as the future interface for smart homes, but most implementations have been limited by language support, rigid command structures, and poor handling of natural speech variations. By supporting eight languages and understanding intent rather than syntax, Havells is addressing a real pain point for Indian households where linguistic diversity is the norm, not the exception.
The partnership between Havells and ElevenLabs also highlights how specialized voice AI companies are becoming critical infrastructure for consumer electronics manufacturers. Rather than building voice capabilities from scratch, companies can now integrate pre-built, production-ready speech models that handle the complexity of multilingual, real-time interaction.
"This Agentic AI Voice capability is a breakthrough for our customers, allowing them to control and interact with their devices naturally, without having to remember commands or learn rules. Its potential is limitless and will continue to expand as new capabilities are added across our devices and services. As a daily-use AI experience, it showcases the transformative potential of AI agents, far beyond the traditional voice interfaces of the past," stated a Havells representative.
Havells Leadership
The feature is now live for all 2.67 million Havells One app users, making it one of the largest deployments of multilingual agentic voice technology in the smart home space. As Havells expands language support and device coverage, the implementation could serve as a template for how other manufacturers approach voice interfaces in multilingual markets.