The Renault Korea booth at NextRise 2026. At first glance, it looks like a showcase for a voice assistant, an in-car game, a taxi app, and 3D navigation. Look a little closer, though, and a different picture emerges. Kakao brought mobility services and autonomous driving data. TMap brought spatial information and routing. SmashLabs brought content. Valeo brought vehicle sensor data. Renault was trying to fold all these different technologies and services into a single user experience inside the car, under one name: "AI Orchestrator." This piece isn't meant to introduce one more AI feature. It's a field report on why Renault Korea is building a separate digital experience for the Korean market instead of adopting a common global platform, and what role Korea is taking on inside Renault Group along the way.

By Sang Min Han _ han@autoelectronics.co.kr
한글로보기






Walk into the Renault Korea booth at COEX Hall A and you lose your sense of direction for a moment. It's a car company's booth, but there are too many screens. Kakao Mobility and Valeo are on the right, TMap Mobility and SmashLabs on the left. Even inside the Filante demo car, gripping the steering wheel, the instrument panel and the large display beside it take over. Then a voice answers, over a purple screen where stars drift by.
"Hi, I'm Renault Korea."
The AI that introduces itself this way is what Renault Korea calls the Orchestrator.
At NextRise 2026, held in Seoul in June, Renault Korea built its booth around a concept called "Symphony of Mobility." I tried the demo on site and talked with staff there. In the end, the point of the exhibit wasn't any single AI feature.







More Than a Voice Command

Here's what actually happened on screen. I said, "Set the temperature to 22 degrees," and the AI replied, "Done. All seats have been set to 22 degrees." The screen just before that still showed a line reading, "Sport mode (Dynamic mode) is already set." At least within that conversation session, it seemed to be carrying over the earlier setting. Then, when I said "Let's go to a nearby mart," the AI brought up a list: "Here are nearby marts. Where would you like directions to?" and displayed two options in order.
On the surface, this doesn't look much different from any voice assistant. But Renault describes it differently. A Renault representative said the system was "designed to go beyond executing single commands, acting as an intelligent layer that can connect vehicle functions, navigation, infotainment features, and external services into one coherent interaction." Renault calls this system an "intelligent layer." Rather than listing individual functions, the idea is to build a separate layer that coordinates vehicle functions and external services.
Renault's choice of the word "layer" wasn't accidental. What exactly runs underneath it, though, is a different question. I asked whether the LLM was built on an external API or a proprietary fine-tuned model, and how processing was split between cloud and on-device. The Renault representative said, "Certain technical choices are Renault Korea's intellectual property and are still under evaluation, so we can't disclose specifics."







The Latency Problem

When I asked for navigation guidance, the word "Thinking..." briefly appeared on the left side of the screen, followed by, "It will take about 8 minutes (1.8km) to Wonilsa. Shall I start guidance?" Those few seconds of silence stuck with me.
Having tried in-car AI in a number of places around the world, response latency is a problem that keeps coming up. When I raised it, a Renault representative said, "Especially in a vehicle environment where attention and safety matter, customers expect responses to feel instant. We can't disclose specific performance targets at this stage, but we're actively evaluating various architectures that balance local and cloud processing." As many requests as possible are meant to be handled on an in-vehicle AI computing platform, with cloud resources used only when necessary. Another Renault Korea staff member at the booth put it more simply: "That's something our team is still looking into."



A Korean Branch Split Off From the Global Platform

What made me think the most on site was the fact that Renault Korea's infotainment system differs from headquarters'. A Renault representative said, "The Korean market has specific requirements, particularly around navigation," adding, "To address these, Renault Korea is developing a specialized infotainment system that allows integration with leading local services."
Most global automakers go the opposite way. Building one platform and applying it worldwide is more efficient in terms of cost and development. Renault Korea, instead, is developing a separate infotainment branch tailored to the Korean market rather than simply adopting the global platform.
Asked about the reasoning, the Renault representative said, "Across every engineering domain, Renault Korea draws on the expertise, technology, and resources of Renault Group R&D, while also having established a 'Software Studio & Digital Lab' within its R&D center, dedicated to developing solutions for Renault Korea vehicles and the Korean market." The AI Orchestrator is the flagship project to come out of that organization. Even so, the language used to describe its standing was careful: what began as a proof of concept has gradually grown into an important pillar of the future infotainment strategy. Rather than a finished strategy, it's something that started as a PoC and has since been given more weight than that.




3D navigation at the TMap Mobility zone


A Grand Koleos-based concept car on display at the Kakao Mobility booth, NextRise 2026. Judging by the lidar and camera package on the roof alone, it could pass for a mapping vehicle, but the passenger display in the back seat and the ride-hailing branding point to a robotaxi concept instead. The gap between the "high-precision mapping" language of the MOU announcement and the "robotaxi" explanation given on site sits inside this one car.



Behind the Taxi App and the 3D Navigation

So why run this kind of experiment in Korea? Renault gave a fairly clear answer to that question. "Operating an R&D center in Giheung and a plant in Busan means we're exposed to one of the world's most demanding customer bases and one of the most dynamic tech ecosystems, where companies like TMap and Kakao keep pushing innovation forward. That environment makes an ideal platform for developing and validating next-generation digital experiences."
The Kakao Mobility demo ran the Kakao T driver app on the Grand Koleos's panoramic screen and added passenger entertainment in the back seat. But it didn't stop there. On the roof sat a lidar and camera sensor package — unusual equipment for a concept car meant to demo a taxi app.
"This is our first partnership with Kakao Mobility around autonomous driving. We'll be working with Kakao Mobility on various areas from L2++ upward. This is a robotaxi concept, and the equipment on the roof is Kakao Mobility's data collection sensors."
The 3D navigation at the TMap Mobility zone told a similar story. On screen, road signs and buildings passed by in rendering as three-dimensional as the actual road, with lane-level guidance. Maps like this haven't been applied in Korea yet. A Renault representative said, "Our software-based PoCs are always developed with a clear goal of assessing mass-production deployment potential within the next 12 months," adding that "public response to TMap was very encouraging."




The SmashLabs rhythm game and Valeo's XR demo are not concepts. Both features are already available in production Filante vehicles.



Content, or Sensors

At the SmashLabs zone, a rhythm game called 'R:Rush,' set to AI-generated music, was being demoed for multiple players at once. It ran on the car's speakers and screen — content, with no connection to driving data or sensors.
Valeo's zone told a different story. 'R:Racing' looked at first like a car racing game, but it was actually an augmented reality display built on Valeo's XR SDK. Vehicles and objects on the road were overlaid in real time with detection boxes — not a game effect, but a direct visualization of what the ADAS camera was actually recognizing.
A Renault representative explained, "It uses the ADAS system's cameras and ultrasonic sensors." In effect, it showed passengers, through the form of a game, what the car's own eyes were seeing. If the lidar at the Kakao booth was collecting autonomous driving data, the camera at the Valeo booth was repurposing an existing safety sensor as entertainment. Where SmashLabs showed what people could do inside the car, Valeo showed what the car itself could see and sense.
Neither is a concept. Both are already running in production Filante vehicles. On the partnerships with Kakao, TMap, Valeo, and SmashLabs overall, the Renault representative said, "The technologies shown at NextRise shouldn't be viewed as isolated projects, but as examples of a broader long-term strategy."



A Digital Experience Coming Together

Walk the booth long enough and you arrive at the question of what role Renault Korea plays within the group. "Within Renault Group, Renault Korea's role is to serve as the hub for D- and E-segment vehicles, for customers who expect best-in-class experiences in infotainment, connectivity, and ADAS."
That line is worth taking with some caution before reading it as "Renault Group's digital hub." The AI Orchestrator, for instance, was developed initially with Renault Korea customers in mind, but its process and results are shared with headquarters, and as with other successful innovations, its elements may be applied more broadly within the group. That's not a commitment to roll out the entire AI Orchestrator as a group standard.
What NextRise showed wasn't a display of a few features. It was the possibility that different technologies and services could expand into a single vehicle experience. Kakao T's taxi app running on the panoramic screen, TMap's 3D navigation rendered as three-dimensionally as an actual road, the AI carrying over a driving mode it had set earlier while adjusting the temperature — all of it was an attempt to combine Korea's maps, content, mobility services, and user needs into one vehicle experience. At the same time, the specific AI model and architecture haven't been decided or disclosed, and the screen still shows "Thinking..." while waiting for a response. Whether the Kakao Mobility partnership extends past an interface demo into some level of actual autonomous driving, and whether the "12-month mass-production assessment" principle Renault Korea describes turns into a real vehicle program, both remain to be seen.
Still, one thing is clear: Renault Korea is no longer just bringing headquarters' systems over to Korea. That's why the AI Orchestrator's significance isn't about a single piece of technology — it's a shift in role. A move toward an organization that builds and tests new digital experiences of its own. The AI Orchestrator was the name that captured that shift most clearly.

AEM(오토모티브일렉트로닉스매거진)