DEMO
acme.com ▾ AG Log out

Brand overview

Last scan: today 6:04 AM · next full sweep: Fri · 45 questions = 180 answers checked
Citation Score
115 of 120 ▲ +2 fixed this week
answers about your brand that are right
How we count ↓
AI recommends you
19 of 60
buying answers where you're the pick · Fernwell: 24
Answers not right
5
3 wrong · 1 outdated · 1 missing · 1 of them verifying
Est. monthly exposure
14.2K
buyers hearing the 5 not-right answers
How we count

Counting rules

Every answer is right or it isn't. Right means consistent with your verified facts. Not-right splits into wrong (contradicts them), outdated (behind a real change), and missing (no usable answer). A verdict only changes after two consecutive scans agree, so one flaky response never moves your numbers.

Where importance lives

Each flag's exposure = the question's estimated monthly volume × the model's share of AI usage. Errors on big questions in big models rise to the top because more buyers hear them. The arithmetic sits on every flag and is always labeled an estimate.

What moved this week

    No composite score, no invented weights, no 78-out-of-100 to take on faith. These numbers move only when an answer actually changes, every change carries a date and the answer that caused it, and question-set changes are versioned in your history.

    Answers right, last 30 days

    Share of the 120 checked answers that are right

    Right answers by model

    Of 30 checked per model, every day

    Flagged answers

    Human-reviewed before they reach this feed. Verdicts change only after two scans agree.

    Who AI recommends

    Top-recommendation counts across 60 buying answers (15 questions × 4 models)
    You (acme.com) Competitors

    Scan activity

    Question set v2

    45 questions: 30 brand facts, 15 category and comparison. Daily delta covers flagged questions plus the top 14 by volume.

      Want more than corrections?

      Grow builds the presence you don't have yet: new citations, content models recommend, visibility where AI actually looks. Scoped to your niche.

      Book a Grow call →

      Fix queue

      Ordered by estimated exposure. Approve a fix and our team executes it.

      Every fix you approve is implemented by our team, joins the daily scan set, and shows as verified after three consecutive clean reads. Prefer to handle one in-house? Mark it done and we verify it the same way.

      Work log

      What our team changed and when. Entries appear the day the work happens; the scans that follow do the verifying.

      Models take time to re-read corrected sources, so a fix can be real days before the answers move. This log is the record of the work itself, independent of the scan cycle.

      Demo data. All numbers are sample values; on a live dashboard every count links to the answers behind it.