Volkswagen of America
genAI owner assistant over a knowledge base (multimodal RAG)
Volkswagen of America integrated Google Gemini (Vertex AI, BigQuery) into its myVW owner app in September 2024, with an assistant that answers in natural language from manuals and guides and identifies dashboard warning lights via the smartphone camera.
Objective
Make ownership information more accessible in the myVW app to strengthen the owner relationship after purchase: answer vehicle questions in natural language and identify dashboard warning lights, without the customer digging through the paper manual.
The deployment
Volkswagen of America integrated Google's generative AI into its myVW owner app to power a virtual assistant. The owner asks questions in natural language (for example, how to change a tire) and the assistant answers using the manuals, FAQs, help articles, official Volkswagen YouTube videos, and step-by-step guides. The multimodal capability also identifies a dashboard warning light by pointing the smartphone camera at it. Announced on September 26, 2024, initially available on the model year 2024 Atlas and Atlas Cross Sport. The setup relies on Google Gemini, Vertex AI, and BigQuery.
Results Proof C
Production deployment documented by established marketing press naming Volkswagen of America and by the brand's press release; limited initial scope (two models, one market) and no public adoption metric, hence level C.
How it works
Documented architectureThe stack in detail
- llm Google Gemini Generative and multimodal model: natural-language answers and identification of dashboard warning lights via the camera.
- plateforme Google Vertex AI Orchestration platform for the RAG over the knowledge base (manuals, FAQs, guides, videos).
- infra Google BigQuery Warehouse that carries the indexing of the content queried by the assistant.
- outil App myVW Volkswagen of America's owner application, the surface of the virtual assistant.
How it runs, concretely
For ops teams-
1Owner's question customer
The customer asks a question in natural language in the app, or points the camera at the dashboard.
-
2Context retrieval AI
The assistant queries the indexed base (manuals, FAQs, help articles, videos, guides) via Vertex AI and BigQuery.
-
3Answer generation AI
Gemini drafts the answer, incorporating image analysis for warning-light identification.
-
4Base update data team
The teams maintain the source documentation to keep the answers accurate as the range evolves.
The owner's question and the indexed knowledge base (manuals, FAQs, guides, videos). If the documentation is not up to date or poorly indexed, the answers drift or become wrong.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- structured product documentation (manuals, FAQs, guides)
- indexable video content and help articles
Org prerequisites
- customer experience / digital team
- knowledge base maintenance process
Possible stack
- Gemini / Vertex AI or a cloud LLM with RAG
- warehouse for indexing (BigQuery or equivalent)
- vision component for multimodal
The plan, step by step
- Step 1Inventory the product documentation (manuals, FAQs, help articles, videos) and assign an owner per source.Deliverable: Complete source corpus with update owners.
- Step 2Index the corpus into RAG and connect the LLM.Deliverable: Sandbox assistant that answers correctly on a set of owner questions.
- Step 3Add the multimodal component (recognition of dashboard warning lights).Deliverable: Camera feature tested on the most common warning lights.
- Step 4Launch the beta in the app on a limited model scope.Deliverable: Feature live with in-app feedback collected.
- Step 5Set up the knowledge base maintenance process and extend to the range.Deliverable: Documented update ritual + model extension plan.
First step: Index the manuals and FAQs into a RAG and connect a natural-language assistant into the owner app.
The same pattern, proven in other industries
ADT
genAI customer agent
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Sources
- S1 Volkswagen integrates Google's generative AI to enhance app experience Established press archive pending
- S2 Volkswagen integrates AI into the myVW mobile app Primary archive pending
- S3 How Volkswagen is utilising AI to enhance its digital assistant Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.