What Sage Is (and Isn't) For

Sage isn't a tool for interviewing AI personas. If you want to interview AI personas, that's the Interview feature. Sage builds an AI expert from your materials, so you can ask that expert questions whenever you want.

This guide is for the times when you ask "can I use Sage to interview personas?" — and the answer is no, plus what Sage is actually built for.

The two features side by side

If you've been thinking of Sage as "a tool to ask AI personas questions" — it's not. It's "a tool to ask an AI expert questions, where the expert learns from materials you give it."

What Sage actually does

You upload documents (PDFs, Markdown, web pages, transcripts — whatever you have). Sage reads them, builds a knowledge base, and lets you ask questions like you'd ask a colleague. The expert gets smarter over time: you can add more materials, edit its knowledge directly, or run "supplementary interviews" where you answer questions to fill in gaps it has identified.

Common use cases:

• Capture expertise before someone leaves. A senior designer's 20 years of decisions become a Sage expert. New hires consult it instead of pulling the designer away from current work.
• Personal study assistant. Upload course notes and papers. Sage helps you review and answer questions in the language of your course, not generic explanations.
• Industry tracker. Drop in research reports and earnings announcements. Ask "what changed this quarter" and get an answer grounded in the materials you provided.
• Domain-specific advisor. Legal, medical, technical — anywhere you have reference materials and need quick, sourced answers.

The flow

1. Create the expert. Upload materials (PDF, Word, Markdown, audio, URLs — up to 10 files per expert). Sage reads them and proposes a short bio, areas of expertise, and suggested starter questions.
2. Ask questions. Treat the conversation like messaging a knowledgeable colleague. Sage answers based on what you uploaded, and cites which material a given answer came from.
3. Fill gaps as they appear. When Sage doesn't know something, it flags the gap. You can upload more material, write in the answer directly, or run a supplementary interview where Sage asks you questions to learn the missing pieces.
4. Keep evolving. As new materials appear (updated specs, new research), add them. The expert's answers update accordingly.

Supplementary interviews, explained

This is the part that confuses people most. A supplementary interview is you being interviewed, not the other way around. Sage asks you questions to extract knowledge it needs. Your answers get added to its knowledge base.

Example: You built a UX design expert from your company's design system docs. The expert notices it's missing mobile specifications. You start a supplementary interview, Sage asks "what are your iOS navigation patterns?" and "what are your Android differences?", you answer in your own words, and the next time someone asks the expert about mobile, it has the answer.

What Sage is not good at

• Market research or user interviews. Sage has no opinion about your customers. Use Interview for that.
• Real-time data queries. Stock prices, weather, today's news — Sage works from materials you gave it, not the live web.
• Replacing ChatGPT for general knowledge. Sage is built for your knowledge base, not the world's.

When to use what

• Want to understand what users want → Interview
• Want to consult an expert advisor on a topic you've uploaded materials about → Sage
• Want to know what's happening in a market → Scout (social media observation) or Research

Related

• How Sage compares to using ChatGPT with a custom persona
• Creating your first interview