After introducing a voice‑based AI assistant, Rivian’s founder and CEO RJ Scaringe told the TechCrunch Disrupt conference this week that the company intends to enable third‑party AI agents — from firms such as Anthropic — to run on its vehicles in the future.
"That’s where it becomes magical," Scaringe said onstage. He did not specify a timeline for when those integrations would be available, and Rivian declined to provide further details, referring instead to its "AI & Autonomy Day" event planned for December.
Rivian’s in‑house agentic framework
Rivian has developed its in‑car software stack from the ground up rather than relying solely on Apple CarPlay or Android Auto, and it has built many of its own applications such as navigation. The company says it created an agentic framework that interprets what action a user wants and routes that request to the appropriate part of the vehicle software.
Scaringe explained that the system acts like a traffic light: when a user asks something, the vehicle decides which model to use. Some requests are handled by local models running at the edge on the vehicle; more complex queries can be forwarded to cloud models. Different models are used for different kinds of tasks.
Existing external app access and plans for frontier AI
As of May of this year, the agent could access external apps like Google Calendar. Scaringe said the next step would be to add products from frontier AI companies — he mentioned examples such as Claude, OpenAI’s Dots, or Meta’s Muse — but cautioned that making them work in a car is nontrivial.
"There’s a lot of plumbing behind the scenes we have to do to make sure that that works, and it’s not just within Rivian. That’s also a lot of coordination with the labs to make sure they enable that plumbing," he said.
Clear safety boundaries: no vehicle‑level control by agents
Scaringe emphasized that any software integration will avoid granting agents control over safety‑critical vehicle functions. "You wouldn’t want an agent to determine, ‘hey, it’s time to apply the brakes, or time to steer the vehicle,’" he said.
Instead, Rivian expects integrations to extend how people already use agentic apps: agents would assist by leveraging user context and past conversations, while the vehicle’s own systems retain control over driving functions.
A user scenario
Scaringe offered a practical example: a driver could ask the Rivian assistant to turn on ventilated seats and at the same time reference a prior conversation about a planned Friday meal with guests. The car could then pull context from an earlier assistant interaction on another device to help with menu planning.
Market context
Rivian would not be the first U.S. automaker to add AI agents to its software: earlier this year, Tesla introduced a feature that let users interact with SpaceX’s Grok chatbot and Grok Bot assistant.
In sum, Rivian is preparing to open its vehicles to third‑party AI agents while stressing that significant engineering work and close coordination with AI labs are required, and that safety‑critical controls will remain outside the authority of such agents. The company’s December event may provide further specifics.
