What is relationship intelligence?

Relationship intelligence is software that works out who knows whom, and how well, from records of things that happened rather than from what anyone typed into a contact list. The records are communications metadata: who emailed whom and how often, who sat in a meeting with whom, who is connected to whom on a social network, whose number is in whose address book. The output is a graph of people with a strength score on every tie, which you can search: who do we know at this company, who has the strongest tie to this person, who can make this introduction.

The phrase comes from private capital. Affinity, which sells to venture and private equity firms at $2,000 to $2,700 per user per year, describes itself as the AI-first private capital CRM and built the category around a fund's email and calendar. The same idea now runs in tools for sales teams, professional services firms and individuals, and the differences between them come down to one question: whose relationships is the graph built from.

Email. The core signal. The systems that do this carefully read headers only: sender, recipients, timestamps, subject. Happenstance reads email as headers and never message content. Reply rates matter more than volume: an unanswered sender is a weak tie whatever the count, which is why Connect The Dots scores each tie on two-way replies.

Calendar. Meetings are a stronger signal than mail, because a shared calendar entry means both people chose to spend the time. Connect The Dots lists calendar as coming soon; the firm platforms and Happenstance read it.

Connections and address books. LinkedIn, Twitter and Instagram connections, and the entries in Google Contacts or Instagram, say that a relationship exists without saying how strong it is. They fill in the people who never emailed you.

Inferred ties. Some platforms add relationships that nobody recorded: two people who overlapped at an employer, or went to the same school. Village's data page labels only 7% of its 4.77 billion edges direct; the rest are co-worker, role and alumni overlaps. Vieu builds its 4 billion node graph primarily from external signals such as shared work history and board affiliations. An inferred tie is a lead to verify, which is a different thing from a relationship.

Finding a warm path is the core job: a seller reaching a buyer through a colleague who knows them, a founder reaching a partner through a founder that partner has backed, a recruiter finding the candidate a teammate worked with, a job seeker finding the former colleague who can refer them. Around that sit meeting preparation (who on our side knows the people in the room), CRM hygiene (the relationships that exist but were never logged), and bringing dormant relationships back into view.

A CRM records what you tell it; relationship intelligence observes what happened. Affinity and 4Degrees fuse the two into one product for a firm. Happenstance skips the CRM: the graph is the product and you search it in plain English. 500,000+ people use it.

A sales database sells access to strangers: LinkedIn Sales Navigator searches LinkedIn's members at $119.99 a seat, Clay waterfalls external data providers, The Swarm sells a 500M-profile dataset beside its pooled networks. Relationship intelligence maps the people you already know from data you own. One finds cold targets and the other finds warm paths, and most teams end up needing both kinds. The category page, relationship intelligence, has the short version, and relationship mapping covers the neighboring term.