Introduction impact
Tending should help an organisation understand whether bringing people and organisations together leads to relationships that continue. This is different from ordinary network inference and needs a stronger evidence model.
The claims Tending can make
There are three deliberately separate layers.
1. Explicit introduction
A team member records that the organisation deliberately introduced two connections. This is stored as a first-class introduction linked to the source moment.
A normal moment that happens to mention two people is not an introduction.
2. Continuation evidence
After the introduction, Tending can look for later retained moments that involve both introduced connections.
That supports a narrow claim:
Activity involving both sides was recorded after the introduction.
It does not prove that the two sides now have a direct relationship, that the introduction caused the later activity, or that the organisation was no longer involved.
3. Independent continuation
The stronger claim — that the relationship continued without the organisation holding it together — must be confirmed by a person who knows that to be true.
Tending records:
- who confirmed it;
- when they confirmed it;
- an optional note or evidence reference.
Automatic continuation evidence can prompt this review, but it never upgrades itself into an independent-impact claim.
Product behaviour
The Network page surfaces an Introductions taking root panel. It shows:
- introductions recorded;
- introductions with later activity involving both sides;
- introductions explicitly confirmed as continuing independently;
- introductions where no later evidence has been recorded yet.
A dedicated recorder lets a team member choose two existing connections, add a note/date, and create both the source moment and the explicit introduction record.
Why this is separate from the graph
Co-mentions are useful for discovering possible network structure. They are intentionally permissive and can arise from one meeting, one note, or one shared context.
Introduction impact is an evaluative evidence trail. It therefore starts with explicit human intent and keeps stronger outcome claims human-confirmed.
The graph can visualise the resulting relationship, but it is not the source of truth for the impact claim.
Future extensions
Useful later additions include:
- recognising introduction language in the moment-understanding model and offering an opt-in suggestion in the composer;
- using retained connected-source evidence to show richer continuation signals while respecting the private-candidate boundary;
- showing introduction outcomes in exports and relationship handover packs;
- connecting introductions to programme/space-level impact reporting;
- recording richer evidence about what happened because of an introduction without treating correlation as causation.