A static assistant makes every user follow the same script, even when their history or preferences are known. An adaptive flow changes the interaction path as context and user behavior change, producing more personal, relevant, and timely responses. section 9.3 section 9.3.1
Design an adaptive flow around the context the assistant can use and the effort it can remove:
- Use previous searches, purchases, preferences, location, or current promotions to tailor recommendations. section 9.3.1 section 9.4
- Anticipate useful follow-up information, such as reviews and comparisons, instead of requiring users to request every detail. section 9.3.1
- Update context when intent changes; a shift from budget travel to luxury accommodations should change subsequent recommendations. section 9.3.3
- Monitor interactions, refresh retrieved information from trusted sources, and refine the flow through feedback and iterative adjustments. section 9.3.3 section 9.4
From generic shopping to relevant guidance
A static assistant asks which laptop category the user wants and lists standard options. An adaptive assistant recognizes that the user previously explored eco-friendly laptops, asks whether that preference still applies, and can add relevant reviews or comparisons. Retrieval-augmented generation (RAG) can supply the response with current contextual information. section 9.3.1 section 9.3
User: Hi, I am looking for a new laptop. Chatbot: Welcome back! Last time you looked at eco-friendly options. Are you still interested in sustainable products?
Compare a generic laptop greeting with a context-aware prompt that uses the user’s earlier interest in sustainable products. section 9.3.1
