Two agents are designed for external users.
One supports vehicle owners during claims. Cheche says it can handle initial loss notification, damage assessment and claims-status updates through channels including vehicle interfaces, customer-service systems and manufacturer apps. The underlying claims platform connects automotive parts catalogues with insurer repair networks.
Cheche says this system is already live with Volkswagen Anhui and Avatr and integrated with PICC, Ping An and China Pacific Insurance.
Another agent supports underwriting and pricing. Insurance professionals can enter a VIN, licence plate or order number and retrieve vehicle-level risk information. The underlying pricing model considers more than 200 risk factors, including vehicle characteristics and data associated with advanced driver-assistance and automated-driving systems.
Cheche says the model is commercially deployed across more than 100 Chinese cities and that the company has agreements with more than 200 insurance companies.
200 insurers show why EV data is becoming infrastructure
Three additional ABAO agents are being used internally for customer-service quality control, settlement follow-up and processing non-standard insurer documents.
Cheche reports a 30% improvement in settlement workflow efficiency and a 50% improvement in overall settlement processing efficiency without additional headcount. Those figures are company-reported and should be treated as evidence from Cheche’s own environment rather than general industry benchmarks.
The wider development matters because electric vehicles create insurance data that conventional vehicles did not generate at the same scale.
Battery information, connected-car systems, advanced driver-assistance data, electronic parts catalogues and manufacturer platforms can all influence pricing or claims.
That creates room for specialised providers.
Smaller insurtech companies do not need to reproduce Cheche’s entire platform. They can focus on one part of the emerging data chain, such as repair-cost intelligence, battery-health evidence, claims documentation, fleet risk, vehicle-data normalisation or integrations between manufacturers and insurers.
A useful first market is likely to be one where electric vehicle adoption is growing faster than insurer software has adapted.
The strongest businesses will probably sit close to proprietary workflow or data access. A generic AI assistant can be copied quickly. A system connected to repair networks, fleet records or manufacturer data becomes harder to replace.
The main risk is access. Vehicle data is controlled by manufacturers, insurers and platform providers, meaning a startup can build an attractive product and still struggle to obtain the data required to make it useful.
For founders entering this market, distribution partnerships may therefore matter as much as the model itself.
More information
Source and more information: Cheche Group Launches ABAO Agent Family, Deploying AI Across the Full NEV Insurance Value Chain