The dashboard is the home screen for a tracked site. It answers a short check-in: how visible you are right now, and the single most useful next step.
What you see first
The hero leads with AI visibility and one of four standings, based on how often answers in the latest trusted full scan named you (your mention rate):
| Standing | Mention rate |
|---|---|
| Growing | Below 40% |
| Contending | 40% to 59% |
| Leading | 60% to 79% |
| Owned | 80% or more |
Growing includes early rates such as 0% and 4%. It is a starting line, not a failing grade. A ladder shows where you sit, and engine rings below it show which tracked engines named you, strongest first.
The hero button usually reads Fix these next and jumps to the Fix These Next card. When setup or a connection needs attention first, it reads Finish setup, a connect or reconnect prompt, or Review latest scan instead.
Under the hero, a row of up to four tiles shows the readings that most need attention: usually Mention rate and Citation rate, plus AI Readiness or a site-health reading while it still has checks left. The Visibility Index card sits at the top of the right column.
Below the tiles, Fix These Next lists this scan's recommendations; the full queue lives on Actions. Beside it are your share of voice and the work you have In progress or Waiting On Evidence. Ask Stride follows, then the AI Engine Overview.
Outcome rates
Start with Mention rate (how often answers named you) and Citation rate (how often answers cited your own site). Each tile shows the count and the number of measured answers behind it, for example 12 of 40. These are the clearest answers to “did the brand show up?” and “did our own pages support the answer?”
How the index is calculated lives behind the info icon beside the Visibility Index (What is the Visibility Index?), not as a methodology paragraph on the page.
Visibility Index
The 0–100 Visibility Index summarizes mention, own-site citation, answer portrayal, and answer position across your tracked prompts. Each measured engine is scored independently and engines are averaged equally.
The index is not a site grade, peer percentile, or competitor rank. Read it only beside the measured outcome rates and within a comparable measurement epoch. See Understanding your Visibility Index for the formula and interpretation rules.
AI Engine Overview
The AI Engine Overview shows the mention rate for each engine your plan includes, with the same Growing, Contending, Leading or Owned standing beside it, using the thresholds above but applied to that engine's answers only. A low rate here is a standing, not a failed health check. Engines your plan doesn't include appear locked, with the upgrade that adds them.
Site-health readings, such as AI Readiness, use a 0–100 health scale. A reading Stride couldn't collect shows — and Not measured yet., never a zero. The full checks live on AI Readiness.
Comparable trends
Dashboard Index trends connect scans with the same normalized prompt-text and engine work set. A competitor-only change does not necessarily restart that Index trend. Action proof uses a stricter measurement epoch that also protects target and competitor evidence, so it can become not comparable even when the Index trend continues.
The dashboard uses the newest trusted scan that covered the full eligible prompt pool when it ran. A later Topic-only baseline does not replace that dashboard reading. When prompts are added, removed, disabled, or paused, the last full-scan dashboard remains visible as a dated snapshot while those changes wait for the next full scan. Priority scans can refresh individual prompt evidence during that wait, but they do not replace dashboard metrics or create a new Visibility Index reading. The next scheduled full sweep advances the dashboard automatically; a manual full scan is optional.
Small or sparse movement stays below the provisional five-point recheck threshold and needs repeated comparable evidence before it is called sustained. That product rule is not a universal noise band or a significance claim.
The Progress card and proof timeline stay off the home view until there is a confirmed from→to on comparable scans. A first scan is a baseline, not a trend.
Actions and proof
Recommendations retain the prompt, answer, source, or technical check that triggered them. The Actions queue uses that evidence, recurrence, connected query demand when available, and implementation effort—never a promised score lift or invented ROI.
After you mark work implemented, its proof timeline shows:
- that implementation was recorded, separately from a passed live check;
- the target prompt, page, or source;
- the expected action-specific signal and waiting period;
- current evidence health, including provider failure, insufficient evidence, epoch mismatch, confounding, integration unavailability, or source deletion;
- comparable observations from later scans; and
- target-aligned connected outcomes only from complete, sufficient 28-day before-and-after windows.
Overlapping Actions are marked confounded instead of each receiving credit for the same movement. Unless a stable control series exists, Stride says a result was “observed after” the work—not caused by it.
Prompt change timeline
Open Dashboard → Prompts → Tracked, then open one prompt. Its Change timeline places three kinds of dated records in one sequence:
- comparable full-scan readings, with the number of measured answers that named your brand and cited your site;
- content drafts saved in Stride for that exact prompt; and
- changes you recorded as shipped from the linked Action.
Drafts are not treated as published work. A shipped change remains Measuring until later comparable scans exist. Associated movement means the measured prompt result followed the change; it does not mean Stride proved the change caused the result. Confounded means another overlapping change prevents separate credit.
Recommendation rank is not shown yet. Recurring scans currently retain answer position, which describes how early your brand appeared in the response and is not the same as rank among recommended brands.
Why is my draft missing? The timeline only uses an exact saved prompt ID or an unambiguous exact prompt snapshot. It does not guess from similar wording.
Why is an older scan missing? Prompt readings stay inside the current comparable measurement epoch. A prompt, engine, competitor, geography, or measurement-contract change can start a new baseline.
Source opportunities
Repeated comparable scans can also surface third-party sources that recur in answers where a competitor appears and your brand does not. These become normal human-owned Actions for expert contributions, original research, video or transcript work, profile completeness, or community participation. Stride does not auto-post, fabricate reviews, promise placement, or claim a source cited a specific competitor when the answer evidence cannot establish that link.
Only open platforms qualify: video and podcast hosts, review directories, communities, maps and local listings, open knowledge bases, and research archives. A competitor's own channel, subreddit, or product page on one of those platforms is skipped. If a suggestion is still wrong, choose Not a fit on the card or in the action and give a reason. Stride stops suggesting that site, and the next-best source takes its place.
Where to find it
Sign in and select a site. The dashboard shows the current measured story; Fix These Next lists that scan’s recommendations; Actions holds the full queue; Analytics holds the implementation proof timeline.