Insights and My coach

Included in the upcoming 1.0 release, pending final native acceptance. These features are included in the existing Plus price; there is no separate coaching add-on. Optional AI provider charges are separate. Free keeps your context notes editable and shows fictional examples of Plus insights without analyzing your real meetings for a preview.

Start with a contact or meeting

Open Contacts, choose a contact, then Overview. Review decisions, actions and included interactions together. Details keeps the existing contact fields and voice controls. In a meeting, open Insights for that meeting’s evidence.

Enable insights deliberately and review Included meetings. Purchasing Plus, enabling voice recognition or recording another meeting does not grant history analysis or cloud permission. Uncertain speaker identities are not joined merely because they have the same display name.

Evidence links open the saved source. Review the original wording and surrounding context before acting. An unresolved action means no completion is recorded; it does not prove the person failed to follow through.

Add Your context

Use Your context for notes such as “My manager,” “Asked for a cost breakdown,” or “Prefers a written agenda.” Choose a category and whether to include the note in insights. An impression is your perspective, not a measured fact about someone. You can edit or delete notes. Do not include unnecessary sensitive details.

Notes stay local by default. Including them in a provider’s insight context needs an additional choice. Personal coaching reflections are kept out of provider prompts and connected-agent results.

Choose a focus in My coach

Choose a professional focus, review a suggestion tied to available evidence, and try one small practice in your next conversation. You can add a private reflection about how it went. The coach may say there is insufficient evidence; it does not have to invent a strength or weakness.

Documentation trends show their periods, meeting counts and denominators. A rise in actions with recorded owners and due information describes the selected notes, not an objective improvement in communication or job performance. Meetings, roles and recording habits can differ between periods. No personality, emotion, confidence, honesty or employee performance score is produced.

AI suggestions can miss context. Review the linked sources and use your judgment. This is professional reflection, not psychological diagnosis or clinical advice.

Keep control of the evidence

Hide an item to simplify a view. Exclude a meeting to remove it from future insight evidence and counts. Restore a hidden item or included source when useful. A forgotten evidence item stays out of generated context until restored; this is not a promise that a model can never express a similar idea from other sources. These controls do not mark actions complete or delete original recordings.

Changing sources or privacy choices invalidates a reflection being generated. The model receives bounded evidence with coverage information. Counts cover the eligible selection, but a suggestion does not necessarily consider every passage. The displayed omissions and unavailable states explain these limits.

Processing and personal agents

On-device AI is the default. To use a provider, approve Insights in AI provider settings and enable cloud insights for the selected history. Optional context sharing is a separate choice. Recordings, voiceprints and acoustic information stay on this device through Markdrip-controlled AI processing.

Existing agent access does not automatically expose this private area. Enable private-insights access separately if you want your connected agent to read the selected evidence. Agent reads are bounded and respect the same exclusions; external agents follow their own data-handling settings.

See Data and network use and Local agent access.