Assignment · Not Started

Project 3: AI Customer Support Assistant

प्रोजेक्ट 3: AI कस्टमर सपोर्ट असिस्टेंट

Design an AI Customer Support Assistant workflow, including an architecture diagram showing: User question → Knowledge source → AI processing → Response → Escalation → Human review. For your design, specify: (1) what the 'knowledge source' is (a fixed FAQ list, a document — connect to why this matters for avoiding hallucination, Module 7), (2) what specific criteria decide whether a query is answered automatically or escalated to a human (connect to Module 9's customer-support automation levels and Module 10's human-approval lesson), (3) what happens if the AI's confidence is low (connect to Module 10's error-handling lesson), and (4) how you would log and periodically review the assistant's performance (connect to Module 10's logging/monitoring lesson). Use a real or realistic business/product context for your examples.

एक AI Customer Support Assistant workflow design करें, एक architecture diagram शामिल करते हुए जो दिखाए: User question → Knowledge source → AI processing → Response → Escalation → Human review। अपने design के लिए, specify करें: (1) 'knowledge source' क्या है (एक fixed FAQ list, एक document — बताएँ यह hallucination से बचने के लिए क्यों matter करता है, Module 7), (2) कौन से specific criteria तय करते हैं कि एक query automatically answer हो या human को escalate हो (Module 9 के customer-support automation levels और Module 10 के human-approval lesson से connect करते हुए), (3) अगर AI का confidence कम हो तो क्या होता है (Module 10 के error-handling lesson से connect करते हुए), और (4) आप assistant के performance को कैसे log और periodically review करेंगे (Module 10 के logging/monitoring lesson से connect करते हुए)। अपने examples के लिए एक real या realistic business/product context इस्तेमाल करें।

Complete architecture diagram matching the 6-stage flow described
A specific, named knowledge source (not just 'the AI knows it')
Explicit escalation criteria and a low-confidence fallback behaviour