Zhikeling CS Agent
AI answers first, humans take the hard ones
Repetitive questions should not consume people. A one-line iframe widget goes live and greets visitors; the AI answers from the knowledge base (it correctly answered a returns-policy question in testing) and hands over when it cannot or the customer insists, with per-channel seat leases and full context carried across shifts. Complex issues become tickets, SLA timing and escalation run automatically, and everything is recorded for review.
Channels include web, WeCom, desktop and mobile with callback protocols; the operations console covers conversations, tickets, agents and merchants. Delivery artefacts pass a secret scan, ship without default passwords, and local verification keys come only from environment variables or files outside the repository.
Note: e-commerce integration is seeded demo data; real marketplace open platforms are not connected yet.
Core capabilities
- One-line embeddable visitor widget
- Knowledge-grounded, traceable answers
- Multi-turn sessions across web, WeCom, desktop and mobile
- Agent workbench with per-channel seat leases
- Tickets, SLA and escalation
- Business metrics and admin console
- Artefacts ship without default passwords, secret-scanned
- On-premises with tenant data isolation
Who uses it
- Pre-sales and after-sales for merchants
- Documentation Q&A for software vendors
- One entry with routing across stores
- Overnight and peak-hour cover
Delivery models
Multi-merchant SaaS · on-premises · source licence · white label
Tech stack
Java 17 + Spring Boot（RuoYi-Plus，含工作流）· Vue3 · Electron · uni-app · MySQL / Redis / RabbitMQ