Design an Agentic Ticket Support System
Design an Agentic AI-Powered Ticket Support System
Design an end-to-end architecture for an AI-powered ticket support system. The system must integrate with two distinct external systems:
- Ticket System: This system displays all support tickets exclusively through a web user interface and does not provide an API. The AI agent must utilize a browser-based automation approach to read and extract ticket data.
- User System: This system offers an API that can be used to retrieve user information. This user data is crucial for generating personalized chat responses.
Your design should comprehensively explain the following aspects:
- How new tickets are discovered.
- How the AI agent reads ticket content from the browser-only ticket system.
- How user context is retrieved from the User System API.
- How the AI generates personalized responses.
- How the system ensures critical non-functional requirements such as reliability, security, and observability.
- How human review is incorporated into the workflow when necessary.
AI interview prep
Core dilemma
Integrating an AI agent with a critical legacy system that lacks an API (requiring unreliable browser automation) while ensuring high reliability, security, and human oversight for sensitive customer support interactions.
3 tradeoffs and 3 stress tests locked
Interview prep
Unlock the interviewer X-Ray rubric, tradeoff matrix, stress tests, clarifying questions, and key concepts.
Design an Agentic Ticket Support System
Design an Agentic AI-Powered Ticket Support System
Design an end-to-end architecture for an AI-powered ticket support system. The system must integrate with two distinct external systems:
- Ticket System: This system displays all support tickets exclusively through a web user interface and does not provide an API. The AI agent must utilize a browser-based automation approach to read and extract ticket data.
- User System: This system offers an API that can be used to retrieve user information. This user data is crucial for generating personalized chat responses.
Your design should comprehensively explain the following aspects:
- How new tickets are discovered.
- How the AI agent reads ticket content from the browser-only ticket system.
- How user context is retrieved from the User System API.
- How the AI generates personalized responses.
- How the system ensures critical non-functional requirements such as reliability, security, and observability.
- How human review is incorporated into the workflow when necessary.