Connect Salesforce and BeforeQuery indexes your Salesforce Knowledge articles, putting your service knowledge base behind every surface the assistant serves. If you also run support in Salesforce, resolved cases can be indexed too.

​
Prerequisites

  • A Salesforce org with Salesforce Knowledge enabled and your instance URL (e.g. https://acme.my.salesforce.com).
  • A connected app with OAuth client credentials (Setup → App Manager → New Connected App) — you’ll need its client ID (consumer key) and client secret, with API read access to Knowledge.
  • A BeforeQuery workspace owner role.

​
Data ingested

Knowledge articles become one document each: the article URL, its title, its summary, and the article body converted from HTML to markdown. Cases become one document each:
  • Case URL in Lightning, and Case 00001234: subject as the title
  • Status, type and priority
  • Description, then every case comment in order
  • The email thread, when you turn it on
A case’s last-modified timestamp is stored, so a case the org reports unchanged is not fetched again.

​
Index Knowledge articles

1

Create a connected app in Salesforce

In Setup → App Manager, create a connected app with OAuth enabled and API scopes that permit reading Knowledge articles. Copy the consumer key (client ID) and consumer secret (client secret).
2

Add the source

In the dashboard, open Knowledge Base → Sources → Add source and choose Salesforce. Enter the instance URL, client ID, and client secret, then click Add.
3

Verify

After the first sync, published articles appear under Documents. Ask an article-covered question in the Playground to confirm citations point at your knowledge base.

​
Resolved cases as knowledge

Cases are their own source type, on the same connected app. Add Salesforce Cases from Add source, enter the same instance URL, client ID and client secret, and it reads Case instead of Knowledge__kav. Each case becomes one document: subject, status, type and priority, the description, and the case comments. Include the email thread appends the email correspondence too — off by default, because signatures, disclaimers and quoted replies are a great deal of text saying nothing retrievable.
Separate from the workspace-level helpdesk integration, which powers copilot drafts and live ticket deflection. You do not need one to index cases.
Cases contain customer PII — enable PII masking before indexing resolved cases.

​
Keeping content in sync

Knowledge articles re-sync every hour, cases every 10 minutes — there is no schedule to configure, and Sync on the source’s page runs one now if you cannot wait. Only changed content is re-embedded. See Sources Overview.

​
Troubleshooting

  • Auth failures? Confirm the connected app is approved for your org, OAuth policies allow client-credential access, and the secret hasn’t been rotated.
  • Articles missing? Only published (not draft/archived) article versions are indexed — check article publication status.