Understand your knowledge corpus
Your knowledge corpus is your organization's curated reference material. It helps supported AI workflows answer questions about your business using relevant information and citations.
Organization administrators manage it in Org settings → Knowledge corpus. The two tabs serve different purposes:
- Raw documents: source material, such as product documentation, pricing guidance, case studies, and extracted file text. These preserve the evidence behind your guidance.
- Derived documents: focused notes that turn those sources into useful guidance for Athean. Each has a title, a description explaining when it is relevant, content, and source citations.
For example, a product specification can be a raw document, while a note explaining how to answer a common buyer question can be a derived document that cites it. Adding source material and preparing the guidance Athean uses are distinct steps.
Add material with the Admin assistant
- Open Org settings → Admin assistant.
- Attach a supported source file, paste the material, or provide a public page URL. Supported file formats include PDF, DOCX, PPTX, TXT, Markdown, CSV, and HTML.
- Explicitly ask to add the material to the knowledge corpus. Attaching a file to a conversation alone does not add it to the corpus.
- Ask Athean to compare it with existing material and propose new or updated derived documents, with citations to the source.
- Review the proposed changes and confirm the work you want it to perform. Check the resulting documents in Knowledge corpus.
Add this product guide to our knowledge corpus. Check for existing coverage, propose concise guidance for common buyer questions, and cite the source for each factual claim. Show me the changes before applying them.
A public URL can be fetched as source material. For a private content library or an ongoing connection to your internal documentation, chat with us about the setup.
Review and edit documents
Review recently changed guidance against its sources before relying on it in outreach or publishing a new snapshot.
- Open Derived documents and use the Last edited column to find revised notes. Both document lists show this column, and document details show the edit date when the document was changed after creation. A blank value does not mean the document has no content.
- Select a note and follow its citation markers to inspect the source material and the passage behind a claim. When available, use the External or Internal source links in the citation evidence or raw-document details to open the original source.
- Read the note's Annotations. Follow Also applies to links to review other guidance affected by the same annotation. This panel displays annotations; it does not provide controls to write them.
- Use Edit and Save changes to update the note, then review the saved content and citations.
If annotations could not be loaded, select Retry annotations. A source link appears only when one is available; you can still inspect the stored source text through the citation. Saving a note and making it available through an active snapshot are separate steps; follow Review which version Athean uses.
To write a note manually, choose New derived document. Add a clear title, a description of when it should be used, and Markdown content. Use Add citation to choose a raw document and provide the verbatim Cited text. Reference that citation in the content and review the rendered preview before creating the document.
To add a PDF yourself, open Raw documents → New raw document. Upload PDF is selected by default. Select Choose PDF, enter a Title (optional) or leave it blank to use the file name, and select Upload and create. Athean extracts the text and opens the raw document. A PDF added this way can also be shared as a hosted page.
For manual source entry, choose Paste text, From URL, or File reference, then add a title and the Extracted text. All three require that text. The URL and file-reference options record where the text came from; use the Admin assistant when you want a page fetched or another file type ingested.
When source facts change, review the derived guidance that cites them. Updating a source does not guarantee that every note based on it has been rewritten. Check dependent citations before deleting source material.
Review which version Athean uses
A snapshot captures a version of the corpus. When your organization has an active snapshot, AI generation uses that published version. Saved edits become available to those workflows after a new snapshot is activated.
If no snapshot is active, generation uses the current saved corpus. Ask the Admin assistant to check which setup your organization uses before treating saved edits as unpublished drafts.
- Ask the Admin assistant to compare your saved changes with the active snapshot.
- Review the changed guidance, source citations, and any warnings about quotes that no longer match their sources.
- Ask it to capture the reviewed corpus as a named snapshot, then explicitly confirm activation when you are ready to make that version available.
The document forms use Save changes; snapshot publication is managed through the Admin assistant. You can also ask it to review and reactivate an earlier snapshot when you need to return to a previous published version.
Check coverage and improve it
Try a real buyer question in a supported workflow, such as Ask knowledge base in Meeting assist, when available. Review the answer and open its citations. Check that the source supports the claim and that the advice reflects how your team wants to respond.
Athean selects relevant derived documents for the question; it does not include every source in every answer. If it reports missing coverage, add or improve the relevant material and check which version is active.
Keep the corpus focused on useful, current guidance: product capabilities, buyer personas, positioning, competitor comparisons, customer proof points, and your sales process. If the knowledge features you need are unavailable, chat with us about your workspace.
See which sources operators use
See which of your sources operators draw on when they write outreach, and which ones they link in their messages. Open Sequences → Content usage. The page is not limited to administrators, and it covers every operator in your organization.
Each row is one generation: the personalized sequence an operator wrote for one contact. Its sources appear by stage:
- Pulled into context: sources the operator read while writing.
- Cited in plan: sources it planned to link. This column is hidden by default.
- Written in outreach: sources linked in the messages it wrote.
Use Add filter to narrow the list by Operator, Source, or Stage reached. Generated sets the period and is always applied. Select a source, the source count, or +N more to see every source the generation used, grouped by stage, with links where available. Select Open draft to review what the operator wrote.
An operator's page in Sequences has a Content usage section that totals each source for the period you choose. Select a count to open the matching generations, or select View generations to see all of that operator's generations for the period.
A source marked no link can be pulled into context but never cited or written, because it has no link to share. To give a PDF source a link, share it as a hosted page for Anyone with the link.
You can also ask Chat:
Which knowledge sources did our operators write into outreach most often this month?