How do you integrate a conversational AI without compromising an existing mobile app?
Diagnosis & challenge
The hello reno iOS and Android app already held a rich history of content on the electric vehicles the carmaker was launching. Adding an LLM assistant, while consistent with Renault's attachment to innovation, risked breaking the coherence of the existing app. The real question was not "show a chatbot" but deciding how it would sit against the existing content, and what tone it could hold on behalf of the brand.
Integrating the LLM layer
I came in to frame the functional design of the reno avatar's LLM layer: how the user reaches it, what interface handles the interaction, what fallback rules apply. As the project went on, I also designed how the avatar handles multimodality: pushing a piece of app content or an external link to create a more unified journey.
Functional evolutions
Beyond the LLM, I supported the client on the app's evolutions to add value and a community dimension: selling R Pass, augmented-reality views of the vehicles, content-page templates for new models, and a calendar of the brand's events.
app downloads
countries deployed
