Do I have to run discovery before my first analysis?

No, but you probably want to. Without discovery, the HubSpot filters at `/new` either don't appear or show raw IDs. After discovery you pick from labels — closed-lost, expansion, demo-scheduled, the real stage names from your portal.

How long does discovery take?

Typically 5-30 seconds depending on the size of your portal. Discovery samples deals to catalog enum values; it doesn't pull every deal record.

When should I re-run discovery?

Re-run after material schema changes in HubSpot — new pipeline stages, renamed deal owners, an extended industry picker. Until you re-run, the filter UI shows the labels from your last discovery snapshot.

Does discovery work with a private app token?

Yes. Discovery runs the same on both OAuth and private-app-token connections. The scope set required for discovery matches the four read-only scopes the integration setup walks you through.

What does HubSpot discovery do, and when should I re-run it?

Integrations / HubSpot·May 31, 2026·2 min·By Callmine Team

Discovery catalogs deal stages, owners, and enum values from HubSpot so the analysis builder can offer real filter options, not IDs.

TL;DR

Discovery scans your HubSpot portal once and catalogs deal stages, owner names, industry values, and company-size buckets. The analysis builder at /new then offers those real labels as filters instead of asking you to type stage IDs. Re-run discovery whenever your HubSpot schema changes.

How it works

When you connect HubSpot for the first time, Callmine runs discovery automatically. It calls HubSpot's API to read pipeline metadata, property definitions, owner records, and a sample of deals — enough to catalog the enum values your portal actually uses. The catalog is stored per-workspace and powers two surfaces:

·The HubSpot filter chips at /new (deal stage, sales rep, industry, company size, amount range).
·The HubSpot filter summary panel that appears on every report so you can see which filters produced the result set.

Discovery never modifies your HubSpot data — it's a read-only catalog build.

Filters discovery enables

Once discovery has run, the analysis builder offers these HubSpot-derived filters:

·Deal stage — picked from your real pipeline stage labels. Multi-select.
·Sales rep — picked from HubSpot owner names. Multi-select.
·Industry — picked from the enum values your portal uses. Multi-select.
·Company size — bucketed by HubSpot's company-size enum.
·Deal amount — numeric range, in your portal's currency.

Combined with Gong-side date range and max_calls, this is the full filter set that determines which calls reach the LLM.

Re-running discovery

Open /settings → Integrations → HubSpot and click Re-run discovery. Useful after:

·Adding or renaming pipeline stages
·Adding new sales reps in HubSpot
·Extending the industry or company-size picker
·Adding a new custom deal property you want to filter on

Until you re-run, the filter UI shows the labels from your last discovery snapshot — your analysis still works, but new HubSpot values won't appear as filter options.