Knowledge graph and stakeholders

The knowledge graph

A network view of everything Auron has extracted about the records you pick. Each record is shown alongside the meetings and chats it was discussed in, the signals detected across those sessions, and the stakeholders mentioned along the way.

You choose what is in it with the entity picker: pick an entity type, then pick the records you care about. Your selection is remembered, so coming back to the screen brings back the same graph.

Working in the graph:

• Click any node to drill into the evidence underneath it.
• Click a source node to focus the graph on that single meeting or chat.
• Double-click a pinned node to release it back into the layout.

The graph is the fastest way to answer "how do these two accounts connect", which is a question no table answers well.

Stakeholders

Stakeholders are the people who come up in conversations: a procurement lead named in a call, a clinician mentioned in a debrief. Auron collects them as they are mentioned rather than requiring anyone to enter them.

Each stakeholder holds a name, a role, an email, an organization, the entity records they are linked to, how many times they have been mentioned, when they were last seen, and free-form notes. Their provenance is kept, so you can see which sessions a stakeholder came from.

You can edit a stakeholder's details, and link stakeholders to each other and to records.

Deduplication

The same person mentioned three different ways produces three stakeholders. Auron proposes merges where it thinks two records are the same person, with a count of how many suggestions are outstanding, and you accept or reject each one. Working through them occasionally keeps mention counts honest.

Both screens link to each other and to Signals from their headers, since the three answer the same question from different angles.