Mahindra Brings Agentic AI Into the Car With Gemini-Powered TEQ_Talk

Mahindra is taking a major step toward conversational in-car AI with TEQ_Talk, a new Google Gemini-powered assistant debuting on the BE 6 SPORTEQ. The integration makes Mahindra the first Indian automaker to deploy a Google Cloud AI agent in a production vehicle.

Built into Mahindra’s MAIA artificial intelligence architecture, TEQ_Talk uses the automotive agent in Gemini Enterprise to control more than 200 vehicle functions. Its capabilities extend across navigation, entertainment, news, climate settings, cabin controls and information from the vehicle’s owner’s manual.

The biggest change is how drivers communicate with the vehicle. Instead of memorizing specific voice commands, users can speak more naturally and allow the AI to interpret their intent. A driver saying that the cabin feels cold, for example, could prompt the system to adjust the air conditioning or close the windows.

TEQ_Talk can also interpret less literal requests. Asking to see the stars could result in the vehicle opening its sunroof rather than simply returning weather information. The approach moves Mahindra’s voice experience away from traditional command-and-response systems toward contextual interaction.

The collaboration also represents Google Cloud’s first automotive AI deployment within an Indian software-defined vehicle. According to Google Cloud India Managing Director Sashikumar Sreedharan, the technology can support more natural interactions across navigation, search, media and communications.

Mahindra is positioning the technology as an evolving platform rather than a fixed feature. The architecture could eventually support additional AI agents designed specifically for drivers and passengers, allowing new capabilities to be introduced as the vehicle’s software ecosystem develops.

For the automotive industry, TEQ_Talk highlights a broader transition in digital cockpits: generative and agentic AI are beginning to replace rigid voice assistants with systems capable of understanding context, intent and multiple vehicle functions simultaneously.