Risk Modifier

What Risk measures

Risk subtracts from your AURA composite score when ChatGPT, Gemini, or Perplexity surface negative coverage when probed. Visually, Risk shows as the orange tint on the AURA ring — the more risk, the darker the ring. Risk does not appear as a separate dimension or ring; it's a modifier.

The intent is to reflect what users actually experience: a brand with high Share of Voice but high Risk (e.g., everyone mentions you because you had a security breach) lands in a worse place than a brand with high Share of Voice and clean coverage.

What counts as Risk

AIVerdict detects Risk via the same AI grounding that drives the AURA dimensions — if AI surfaces controversy when probed, Risk increases. Categories tracked:

• gavel Lawsuits and regulatory actions — class actions, FTC / SEC / equivalents, settlements with material consequences.
• security Security incidents — breaches, leaks, ransomware events, public disclosures of CVE patches with widespread impact.
• person_remove Executive scandals / departures under pressure — reported in the press, especially if the press frames them negatively.
• work_off Layoffs covered as crisis events — routine layoffs are NOT counted; only those covered with negative framing (e.g., "Acme misses earnings, lays off 40%").
• crisis_alert Product safety / outage recurrences — sustained outages, recalls, or product failures covered in industry press.
• trending_down High-volume negative review cycles — review platform pile-ons (e.g., G2 / Trustpilot rating drops 1+ star in <30 days).

How time decay works

Risk is time-decayed because AI training/recall fades old controversies. The current date is computed dynamically per audit so the decay always reflects "now."

• schedule 0–90 days old — full weight (1.0×)
• schedule 90 days – 1 year — half weight (0.5×)
• schedule 1 – 3 years — quarter weight (0.25×)
• schedule 3+ years — near zero (0.05×)

A 5-year-old breach won't tank your score; a 3-month-old breach will. This reflects how AI actually weighs information when grounded.

How to reduce Risk score

• timer Time is the biggest lever. Recent incidents fade. If you had a breach 9 months ago, the score gets better at month 12 (decay step) without you doing anything.
• edit_note Publish post-incident write-ups. A clear "what happened, what we did, what's now different" post-mortem flips the AI narrative from "Acme had a breach" to "Acme handled their breach transparently." We weight transparent post-mortems as partial offset.
• campaign Push positive coverage to dilute the signal. Risk is relative to the volume of mentions. More positive press = lower fraction of negative coverage = lower Risk weight.
• block Don't try to suppress negative reviews. Suppression is detectable (sudden review-removal patterns) and AI grounding finds the suppressed sources anyway via cached / archived versions. Honest engagement beats suppression.

Frequently asked questions

• Is Risk a 6th AURA dimension?expand_more
  No. Risk is a modifier — a subtractor that reduces your AURA composite score when AI surfaces controversies. Visually it shows as the orange tint on the AURA ring rather than as its own ring or dim. The AURA ring still has 5 dimensions; Risk just makes the ring darker if you have negative coverage.
• What counts as Risk?expand_more
  Lawsuits, security incidents (breaches, leaks), regulatory actions, layoffs covered as crisis events, executive scandals, product safety issues, and high-volume negative review cycles. AIVerdict detects these via the same AI grounding that drives the AURA dimensions — if AI surfaces controversy when probed, Risk increases.
• How is Risk time-decayed?expand_more
  Recent incidents (within 90 days) weigh full strength. Incidents 90 days to 1 year old weigh half. Incidents 1–3 years old weigh quarter. Older than 3 years, near zero. The decay reflects how AI training/recall actually works — old controversies fade from training data and live grounding alike. The current date is computed dynamically per audit.
• Will negative reviews always count as Risk?expand_more
  Only sustained negative-review cycles (rapid star-rating drops, review-platform pile-ons) count. Routine negative reviews mixed with positive ones don't move Risk — they belong in Alignment if they accurately describe your product, or just in the broader review distribution. Risk is for crisis-pattern signals, not for normal product critique.