Google is turning Waymo’s newest robotaxi into something more than a driverless vehicle: the Ojai is now also a place where passengers can talk to Gemini, adjust the cabin, and ask questions about the world passing outside the windows. The feature is a small but telling shift in the robotaxi race – one that treats the autonomous car not just as a transportation product, but as a software-defined space for the time between pickup and drop-off.

For years, the most striking thing about a Waymo ride was the empty driver’s seat. You climbed into a modified Jaguar I-PACE, watched the steering wheel turn itself, and experienced an unusual kind of quiet: a car doing all the work without a human at the controls.

Waymo’s new Ojai fleet changes the premise. The all-electric, minivan-like vehicle was designed specifically for ride-hailing rather than adapted from a consumer car, and it has been introduced first to select riders in San Francisco, Los Angeles, and Phoenix as Waymo collects feedback. It is also the debut vehicle for Waymo’s sixth-generation autonomous driving system.

Now Gemini is moving into the cabin. Press an icon on the display and Google’s AI assistant can handle basic environmental controls, answer general questions, explain Waymo, offer trip information, and help a rider explore the area. Someone feeling warm can ask it to set the A/C to 65 degrees; someone curious about a landmark or looking for coffee can ask it to identify places nearby.

That sounds familiar to anyone who has used voice assistants in a personal car. But the context is different. A robotaxi passenger has no steering wheel, no dashboard to manage, and often no personal connection to the vehicle. Gemini is meant to fill some of that gap – not by pretending to be a driver, but by becoming a lightweight, optional layer of help inside a shared vehicle.

A strict dividing line

The most important part of Waymo’s announcement may be what Gemini cannot do. The assistant operates independently of the Waymo Driver, the autonomous system responsible for sensing the world, navigating streets, and controlling the vehicle. Gemini does not control driving or routing, and it has no access to real-time driving data.

Passengers can use a voice request to ask for a pullover, but that appears to be the boundary between the conversational interface and vehicle action. It is a deliberate separation: Gemini can be useful in the cabin without becoming another decision-maker in a system where errors carry physical consequences.

That distinction matters because “AI in the car” can mean two very different things. In one case, an assistant controls playlists, climate settings, or local search. In the other, AI is entrusted with the split-second, safety-critical task of operating a vehicle. Waymo is keeping those jobs separate, at least for now, and framing Gemini as “a helpful passenger,” not as a co-pilot.

It is also a practical product decision. If a passenger asks a messy, open-ended question, Gemini can make a mistake without putting the ride at risk. The driving stack, meanwhile, remains its own specialized system – one designed around perception, prediction, and motion planning rather than conversation.

The cabin becomes software

Gemini’s arrival comes alongside a redesigned in-car interface, Waymo’s first major cabin design refresh in years. The Ojai uses three large displays, but they are not simply mirrors of one another. Waymo says the system adapts to which seats are occupied, presenting fuller controls where they are needed while showing simpler trip-status or media views elsewhere.

That is a more thoughtful approach than treating the rear cabin as a row of identical screens. A lone rider in the right rear seat may get the full set of controls, while other displays fade into the background. In a shared ride, the setup can serve multiple passengers at once. Waymo has also added a Calm Mode that dims the screen and reduces visible trip information for riders who want a less stimulating journey.

The design reflects a broader truth about autonomous transport: removing the driver changes the car’s interior priorities. The cabin no longer has to revolve around one person watching the road. It can be a quiet place to work, a compact social space, or simply a room where people want the technology to disappear until they need it.

Waymo calls the Ojai a “living room on wheels,” which is the sort of phrase automakers have been using for years. But the company is pairing the language with concrete design choices: a ride-hailing-first layout, climate and media controls distributed across the cabin, and a conversational assistant that is available on demand rather than constantly demanding attention.

The useful version of in-car AI

The best case for Gemini in Waymo is not that it makes every trip chatty. Most people will not need an AI assistant to get across town. The value is in those small moments when touching a screen is inconvenient or when a rider wants local context without pulling out a phone.

A tourist might ask what they are looking at. A parent might adjust the cabin temperature while managing bags and children. A business traveler might want to know how far away they are from a meeting. In those situations, voice interaction can feel less like a novelty and more like a natural convenience.

There is a potential downside, too. In-car assistants can become another source of friction if they mishear passengers, respond too slowly, or give overly confident but unreliable local recommendations. Waymo is wisely describing Gemini in Waymo as a beta product and says it expects to extend its voice-assistant capabilities over time.

Privacy will be just as important as usefulness. Waymo says Gemini is inactive until a passenger chooses to engage with it, an opt-in posture that makes sense in a vehicle used by many unrelated riders throughout the day. In a personal car, an always-ready assistant may be an accepted trade-off. In a robotaxi, where passengers may be discussing work, family, or travel plans, it can feel much more intrusive.

A Google ecosystem on wheels

The Gemini integration also highlights a strategic advantage Waymo has that few robotaxi competitors can match: it sits inside Alphabet. Waymo brings the autonomous driving operation; Google brings a mature AI platform, local-search knowledge, mapping infrastructure, and a familiar consumer brand.

That does not automatically make the experience better. A capable assistant still has to earn trust in the noisy, unpredictable setting of a moving vehicle. But it does let Waymo experiment with an experience that rivals will struggle to reproduce with the same degree of vertical integration.

The timing is significant. Waymo says its driverless service has completed more than 20 million fully autonomous trips across more than 11 cities, while the Ojai is positioned as the vehicle intended to help the company expand its footprint further. As the service spreads, differentiation will matter. Riders will compare wait times, prices, coverage areas, vehicle comfort, and whether the ride simply feels easy.

Gemini will not be the reason most people take a Waymo. They will book because they need to go somewhere. But if Google and Waymo get the details right, it could make the autonomous ride feel less like a technology demo and more like a complete transportation experience: a car that drives itself, a cabin that responds to its occupants, and an assistant that stays quiet until it is genuinely useful.