AI Customer Service
AI customer service is used to learn product descriptions, FAQs, customer service documents and provide automated Q&A services to customers or employees.
First task: validate an answer to a customer question
Prepare: Ask the support owner to identify a configured service agent and an FAQ approved for the intended audience, and confirm that the material has finished loading/indexing. Test internally before external use.
- Open Zhichao AI > My AI Tools (Agents) (我的AI工具（智能体）) and select the assigned service agent. “BabelBird Online Customer Service” in the screenshot is an example, not evidence that your organization has an agent with that name.
- Ask a question covered by the FAQ, such as “Which documents are needed for an after-sales request?”, then ask one that the FAQ does not cover.
- Compare each answer with the approved FAQ version. Open citations where provided. Have the support owner check service commitments, prices and policies, and confirm that unknown answers are identified as unconfirmed.
- Completion check: Known answers match the approved material, unknown questions do not produce invented commitments, and users can find the organization's designated human support channel. The organization must supply that channel; this does not establish an automatic handoff feature.
If the entry is missing or answers are inaccurate: Ask the configuration owner to check agent access, knowledge scope, source version and loading/indexing status. For a new agent, prepare the requirements described in Custom Agents and Knowledge-Base Bots.
Choose the publication scope when publishing the service agent. People within that scope can view it, and anonymous access is supported. Its knowledge scope comes from the files selected in the BabelBird directory or newly uploaded. Before external publication, check that the selected knowledge suits the audience, then test with an intended account or an anonymous visitor. Follow the delivered configuration for website embedding. Passing the internal trial alone does not verify external access.
Function description
AI customer service is Zhichao AI’s module for external service and internal support. Enterprises can provide product manuals, FAQs, customer service documents, after-sales procedures, policy descriptions and other materials to allow AI to learn and provide conversational support.
Main abilities
- Answer customer or employee questions based on company-approved information.
- Access to corporate website, help center or business portal through embedded code.
- Can be used as a portal for pre-sales Q&A, after-sales support, product descriptions, internal IT support or knowledge base.
- Supports improving answer stability through fine-tuning, prompts, question and answer range and manual review process.
- It can reduce the labor cost of repeated questions and answers and transfer complex questions to human support or business personnel.
Data preparation
It is recommended that enterprises prepare information with clear structure, clear version, that can be disclosed or authorized for use, including product descriptions, FAQs, terms of service, operation manuals, troubleshooting, price specifications, delivery instructions, etc. In particular, external customer service must confirm which content can be answered publicly and which questions must be referred to a person.
Permissions and usage boundaries
For shared requirements on source authorization, model services and human review, see Zhichao AI Overview. The completion checks for this task do not replace your organization’s access approvals.