A relationship score is a derived signal that estimates the strength of a person’s connection to another contact or account. It can help teams decide who should make an introduction without revealing the underlying conversation. Used carefully, it supports collaboration while preserving the confidentiality of the original source data.
What a relationship score measures
A relationship score is a derived signal that estimates the strength of a person’s connection to another contact or account. It is not the underlying communication itself, but a summary of observed interactions, affinity, or trust cues that can help teams prioritise introductions, routing, or relationship ownership. The score is useful only when the inputs, purpose, and privacy boundaries are clearly defined.
How relationship scores are built
Relationship scores usually combine multiple signals, such as communication frequency, reciprocity, shared context, and recency. The important design choice is whether the model is meant to reflect a durable relationship or just a short-term interaction pattern. If those distinctions are not explicit, the score can overstate closeness, understate weak ties, or produce misleading recommendations.
Because the score is derived, its value depends on data quality and feature selection. A system that overweights one channel, one team, or one kind of activity can produce biased results even when the maths is correct. Good implementations therefore treat the score as an indicator, not a fact, and keep the scoring logic explainable enough for reviewers to understand why two people were ranked as closely connected.
Why it matters for collaboration and privacy
The main benefit of a relationship score is that it can help people choose the right introducer, account owner, or internal contact without exposing the source conversation or every raw interaction. That makes it a privacy-preserving coordination tool rather than a surveillance feature. Used carefully, it supports collaboration while reducing unnecessary disclosure of sensitive relationship data.
The same design can also create trust problems if it is used beyond its intended context. A score intended to guide introductions may be mistaken for a proxy for approval, influence, or authority. When that happens, the organisation risks over-automating human judgement and giving the score more weight than the underlying evidence can support.
Where relationship scores break down
Relationship scores are weakest when the source signals are sparse, noisy, or unevenly distributed across teams and channels. They can also become stale quickly if the model does not decay older interactions or reflect changes in ownership. In practice, a score that looks precise may simply be capturing convenience, email volume, or organisational hierarchy rather than a meaningful relationship.
Another failure mode is privacy leakage through inference. Even if the raw messages stay hidden, the score itself can still reveal that two parties interact frequently or that a specific account has a strong connection to a sensitive contact. That makes access control, purpose limitation, and careful audience design part of the score’s security posture.
Practitioner Guidance
Why practitioners should care: Relationship scores are best treated as decision support, not as ground truth. The safest implementations define what the score is allowed to influence, keep the scoring inputs narrowly scoped, and prevent the output from becoming a backdoor into private communications or confidential source data.
What to watch for: Be cautious when teams start using the score for decisions it was never designed to make, such as trust evaluation, performance review, or access expansion. That is usually the point where a useful collaboration signal turns into a governance problem.
Related resources from NHI Mgmt Group
- When should organisations escalate a high-risk identity score?
- Who is accountable for third-party access when a vendor relationship ends?
- What is the best way to score AI agent workflows in production-like environments?
- How should security teams handle third-party NHI access that outlives the vendor relationship?
Deepen Your Knowledge
Free weekly newsletter
Subscribe to the NHI & AI Identity Journal
The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.
Bonus 33% off our NHI Course when you subscribe.
Reviewed and updated by the NHIMG editorial team on October 30, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org