Intelligence
One surface with six lenses: Intelligence, Ground, Where you lose, Engines, Gaps and Explore, plus the forecast record. Scale and Enterprise.
Intelligence is the analysis surface. It is named after the question it asks, and the other five lenses on it are the evidence for that question. Sources and the forecast board were separate pages until they merged here, and both their links still land.
What you'll learn
- The six lenses on the surface and what each is for
- What the forecast board shows and how its accuracy is tracked
- Which plans include it
The six lenses
| Lens | What it shows |
|---|---|
| Intelligence | The verdict: where you stand and what is driving it |
| Ground | The counted source graph underneath everything else |
| Where you lose | The contested prompts, and to whom |
| Engines | Per-surface behaviour |
| Gaps | What is missing that could be won |
| Explore | Free movement through the same counted graph |
The surface is figure-first: a verdict band at the top, then the lenses that justify it. You can descend from a verdict to the rows behind it without leaving the surface.
The forecast board
The same surface hosts the forecast record: where your citation rate is projected to go, and, more importantly, what happened the last time a projection was made.
- A forecast is a range, not a number, and the range is the honest part.
- Past runs are kept as simulations you can open, so a projection can be read against what actually happened.
- A projection is only scored where a baseline and an outcome exist on the same metric.
See forecasts and ranges for how to read one and what not to promise from it.
Deep links keep working
Two addresses share this surface: one for a source's ladder and one for a past forecast run. Both are carried through the redirect from the pages they used to live on, so a link a colleague sent you last month lands on the right lens at the right depth.
Which plans include it
The Intelligence layer, along with Trace and Industry Pulse, is a Scale and Enterprise feature. On a smaller plan the surface tells you which plan includes it rather than showing an empty board. See plans and limits.
Try this in Outercite
Open Intelligence and move from the verdict band down into Where you lose. Every row there should be a prompt you recognise, and if it is not, your prompt set needs work before the analysis is worth much.
Related
Sources
The Ground lens in detail: the counted citation graph.
Forecasts and ranges
How to read a projection, and what its calibration record means.
Industry Pulse
The cross-customer benchmark, and what it is not.
Last updated on
The last run
What the agents did overnight: the steps they actually took, what they found, and what they wrote, with each part labelled by how solid it is.
Trace
Who arrived from an AI answer, which assistant sent them, and what they were reading. The only lens that connects a citation to a real visit.