Skip to content
Expert Creation — querying a custom expert built from user-provided sources, with retrieved source-passage context shown alongside the answer.

Demo · 02

Expert Creation

Build an AI expert from your own sources — ask it questions, or have it join a discussion.

Expert Creation lets you build a custom AI expert from any body of source material you choose — published scientific papers, industry reports, books, your own internal documents, project notes, or any mix of the above. You walk through a short wizard (identity, voice, sources, a quick test) and the expert is ready to use. From there you can ask it questions directly, or invite it onto a panel in an Experts Discussion alongside other experts from our roster or your own.

Sign up + tryFree credits at signup

Your data is account-tied, not used to train models, deletable on request. See /data-handling.

How this demo works

Approach
A short wizard that turns your supplied sources into a queryable expert.
Deterministic vs generative
When you ask the expert a question, the relevant passages are pulled from the sources you supplied; the expert then answers in the voice you configured. The substance comes from your sources; the voice and framing come from the persona settings you chose during creation.
Human review fit
You don't need to trust the expert blindly — the wizard's Testing step lets you preview responses before you finalise, and every answer afterward surfaces the source passages it drew on, so you can audit any claim back to its source.
See how this fits the broader methodology

How this fits the broader methodology → /how-it-works

Want custom experts and discussions built into your team's workflow — scoped to your data, your sources, and your processes?

Start a pilot