Havells has introduced a multilingual Agentic AI Voice feature for its connected home ecosystem, allowing users to control compatible appliances through natural-language conversations in English and seven Indian languages.
The feature has been integrated into the Havells One mobile app and developed by the Havells Center for Research & Innovation, or CRI. It uses ElevenLabs' real-time speech-to-text and text-to-speech technology as the voice layer, combined with Havells' own conversational intelligence and device-control agents.
The voice feature supports English, Hindi, Tamil, Telugu, Bengali, Kannada, Gujarati and Malayalam. According to the companies, it is available across the Havells One user base, with the app recording approximately 2.7 million downloads.
Havells' earlier voice interface supported English and Hindi on selected appliances but relied largely on predefined commands. The new system is designed to interpret user intent and conversational context instead of requiring specific phrases. It can also process code-mixed speech, a common pattern among multilingual Indian households.
Users can control appliances by device, room or scene through voice interactions. Compatible categories include fans, air purifiers, water purifiers, water heaters, Lloyd air conditioners and washing machines, as well as Crabtree smart switches and other connected products.
The system can also coordinate actions across multiple devices. For instance, a broader request related to room comfort can trigger adjustments across connected appliances rather than requiring separate instructions for each product.
Havells said the AI agent can retain context across multiple turns in a conversation. It can ask users for clarification when an instruction is ambiguous, provide spoken feedback about completed actions and explain certain device error codes in simpler language.
The company evaluated voice AI platforms based on factors including language-recognition accuracy, multilingual capabilities, response latency and voice synthesis. Havells said its CRI team moved the system from initial integration to a production-ready rollout in approximately six weeks, with ElevenLabs providing engineering support for voice quality and regional-language pronunciation.
Havells plans to expand the system with support for more Indian languages and additional connected appliances. The company is also exploring deeper personalisation based on individual household preferences and extending voice interactions beyond appliance controls to functions such as product FAQs, troubleshooting, automation routines and scheduling.
The rollout reflects a broader shift in connected-home interfaces from fixed voice commands towards conversational AI systems capable of interpreting context, intent and multiple languages while coordinating actions across devices.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.