Intelligence

Once a journey has accumulated three or four engagements, Runwita can do more than show you a timeline. The Intelligence layer runs a set of explicit analyses that sit above the raw engagement data: deal health, stakeholder maps, sentiment, blockers, gaps, executive summaries. These are deliberate features, not background magic. You trigger them from the journey page, and they run on the Frontier tier. See AI tiers for which model serves which workload.

​

Why a separate layer

Engagements answer “what was discussed?”. Topics answer “what threads carry across discussions?”. Intelligence answers a different shape of question:

• Is this deal healthy?
• Who in this room actually has influence, and who’s been quiet?
• What objections have been raised that we never came back to?
• Have we made commitments that haven’t been honoured?
• Should I be worried about this journey?

The answers depend on reading across the whole journey, not any single engagement. That’s why they live one layer up.

​

What the layer covers

​

How to run them

There’s one button that runs everything: Run all on the journey’s intelligence panel. It executes all of the above in sequence, sharing context across calls so downstream analyses (deal health) see fresh inputs from upstream ones (stakeholders, objections). Individual analyses can also be triggered one at a time from the same panel if you only want to refresh one. If you haven’t configured the Frontier tier yet, the run errors out cleanly with a message pointing you back to Settings. See AI provider not configured.

​

Where each output surfaces

• Executive summary sits at the top of the journey landing page once it’s been generated. It’s the first thing you see when you open a journey.
• Deal health appears as a coloured chip on the journey header.
• Deal stage shows as a label on the journey header.
• Stakeholder analysis lives in the journey’s people panel.
• Sentiment overlays the engagement timeline as a per-engagement chip and a trend line at the top of the journey.
• Objections and commitment gaps surface in the journey’s right rail as watch items.
• Stale-journey flags appear as a banner on the journey when triggered.
• Meeting brief is generated on demand before a specific upcoming engagement, surfaced separately.

​

When the layer goes stale

Intelligence outputs are pinned to the engagements that produced them. When a new engagement is saved on a journey, all relevant intelligence rows are marked stale. Stale doesn’t mean wrong, it means “not yet refreshed against the new engagement”. The journey UI flags stale outputs visually so you know not to over-trust them. Hit Run all to refresh, or refresh individual analyses. When Runwita launches and notices stale intelligence on any journey (which happens when an engagement was pushed via the MCP while the app was closed), it kicks off a catch-up run in the background after a 30-second grace period. You don’t have to ask.

​

When NOT to run intelligence

Three cases:

• First two engagements. With less than three engagements, the analyses don’t have enough signal to be useful. Wait until the journey has some history.
• Personal journals. If the journey is just you talking to yourself, stakeholder analysis and deal health are theatre. The summary and stale-flagging still help; the rest doesn’t.
• One-off conversations. Standalone events that won’t recur don’t accrue the kind of structure intelligence reads. Save the engagement and move on.

​

What’s next