Baidu AI Cloud Automotive China Translated from Chinese

NIO

Customer Service Automation

NIO wanted to bring large-model capabilities into its smart cockpit so drivers could get richer open-domain Q&A, travel guidance and emotional interaction without sacrificing the in-car user experience.

The first release was pushed to nearly 200,000 vehicles.
The system can choose among more than 16.7 million ambient-light combinations.

Reported impact

Business outcome

NIO launched a more capable in-car AI partner that broadened cockpit scenarios, deepened emotional interaction and showed how large models can differentiate the smart-cabin experience.

  • Customer Experience
  • Scalability

Solution

NIO and Baidu AI Cloud used ERNIE and Baidu's model infrastructure to turn NOMI into a cloud-connected cockpit AI partner across the NT2.0 and 2.2 vehicle platforms. The architecture adds a new NOMI GPT interface layer that can render generated content more naturally, while the large-model service underneath supports encyclopedia Q&A, travel guidance, creative learning, unlimited chatting, expressive mood interactions and vehicle-use questions. One especially cockpit-specific feature is the 'magic atmosphere' workflow, which uses large-model reasoning to map user intents into millions of ambient-light combinations so the car can translate language into a physical cabin effect. The system also uses Baidu's heterogeneous-compute and large-model stack to keep the assistant responsive at vehicle scale. In practice, driver or passenger prompts enter NOMI, pass through cloud-side inference and retrieval logic, and return as multimodal responses inside the cabin interface.

Data flow

User prompts inside the vehicle enter the NOMI interface, are sent to the cloud large-model stack for inference and reasoning, and then return as encyclopedic answers, trip guidance, ambient settings or vehicle-help responses inside the cabin UI.

Solution architecture

4 components · 2 layers
  1. Compute
    • BaiGe Heterogeneous Compute Platform Provides the compute backbone for large-model services behind the cockpit assistant.
    • Qianfan Large Model Platform Hosts and manages the model capabilities integrated into NIO's cockpit assistant.
    • ERNIE Large Model Supplies the reasoning and generation used for Q&A, guidance and emotional interaction.
  2. Serving
    • NOMI GPT Acts as the user-facing intelligent cockpit partner with a new generative interface.

Architecture clues

  • Ambient-light orchestration is driven by language reasoning rather than fixed presets
  • The assistant combines open-domain Q&A with vehicle-specific knowledge and cabin control
  • The feature launched across NIO's NT2.0 and 2.2 vehicle platforms

Evidence from the source

Baidu AI Cloud and NIO built a cockpit AI partner with ERNIE and launched it across all NT2.0 and 2.2 vehicles, with nearly 200,000 cars already pushed (original: 借助文心大模型共同打造蔚来座舱AI伙伴,首批上线蔚来NT2.0/2.2平台所有车型,目前已推送近20万台车).
NOMI GPT supports encyclopedia Q&A, travel planning, learning and creation, unlimited chat and vehicle-use Q&A (original: 百科问答……出行攻略……学习创作……无限趣聊……用车问答).
The assistant can choose suitable ambient lighting from more than 16.7 million combinations based on user intent (original: 在1670多万种氛围灯搭配中,找到适合用户表达的氛围搭配).
Transparency Evidence & methodology 71/100 reference value · 82% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value71/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth25/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity16/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact0/25
    0 metrics describe measurable impact.
  • Deployment scale10/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence82%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence15/20
    3 direct source quotes retained.
  • Date provenance10/10
    The publication date comes from the source page or publisher.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity8/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. Baidu AI Cloud

    NIO customer story

    First-party
    https://cloud.baidu.com/customer/case/weilai.html
    Source type
    Vendor-published customer story
    Published
    May 14, 2025
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date

Date evidence

  • Published on source page May 14, 2025 · Baidu AI Cloud · First-party
    Open date source
Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.