All answers
How it works

Where does relationship intelligence data come from?

From four sources, and none of them is a purchased contact list. Email and calendar metadata show who actually talks to whom and how often. CRM activity shows what the team chose to record. Call recordings add who was in the room. Public employment and education history shows who overlapped before anyone joined your company. The graph is inferred from these, not bought.
October 2, 2026

The four inputs

SourceWhat it uniquely answers
Email and calendar metadataHow often two people actually interact, in which direction, and how recently
CRM activityWhich relationships the company has formally recorded, and who owns them
Call recordsWho was genuinely in the room, as opposed to on the invitation
Public employment and education historyWho worked together before either of them joined your company

The fourth is the one that surprises people. It is how a tool finds that your VP of Engineering spent four years at the same company as a target account's CTO - a relationship that appears in no system you own.

Metadata, not message content

For the email and calendar sources, the useful signal is who, when and how often - not what was said. Frequency, recency and whether replies go both ways are enough to score a relationship, and they carry far less sensitivity than message bodies.

This is worth asking any vendor directly, because the answer varies and it determines what your security review will look like.

What Orbb reads, specifically

Orbb connects to Google Workspace and Office 365 for inbox and calendar, Salesforce and HubSpot for CRM, Gong for calls, and Slack, Teams and Outreach for the rest of the communication surface.

The Google Workspace connection is read-only by design - the documented integration has no capability to write or modify. Gong is connected through OAuth with read scopes for call metadata, participants and transcripts.

Why no vendor can sell you this as a dataset

Three of the four sources are private to your company. Nobody can sell you your own team's interaction history, and nobody can sell you the fact that two of your employees used to work with someone.

That is the structural reason relationship intelligence is a different purchase from enrichment. Clay, ZoomInfo and Apollo sell you records about people. This category computes something that only exists inside your own organisation - which is also why coverage varies so much between companies, and why a trial on your own accounts is the only honest evaluation.

The question that reveals quality

Ask how the tool decides a relationship is strong rather than merely present.

LinkedIn can tell you two people are connected. That is worth very little - a conference handshake in 2016 looks identical to a four-year working relationship. What is worth paying for is the scoring, and the scoring is only as good as the signals above.

Orbb finds the warm paths into your target accounts and names the colleague who can make the introduction.
See it on your own accounts