The questions your buyers ask, put to five AI engines on every run. You get a visibility index, your rank against competitors, the sources behind each answer and a plan to move up.
Tracked on
lumendesk.com
Example · dailyVisibility index
+6 this week
Rank #2 of 4
By engine
Against competitors
01One run
Most prompts never name you, so the score reflects real discovery. Branded and head-to-head questions are tracked too, and reported apart.
Named in 17 of 30 answers
| Buyer prompt | ChatGPT | Gemini | Perplexity | Claude | AI Overviews |
|---|---|---|---|---|---|
| LearnHow do small SaaS teams handle support tickets? | Named | Not named | Named | Not named | Not named |
| ConsiderBest help desk software for a 20-person team | Named | Not named | Named | Named | Not named |
| ConsiderCheaper alternatives to HelpHive | Not named | Not named | Named | Not named | Not named |
| PurchaseHelp desk with per-seat pricing under $20 | Named | Not named | Not named | Named | Not named |
| PurchaseWhich help desk can we set up in a day? | Named | Named | Named | Named | Not named |
| VersusHelpHive vs Lumen Desk for a shared inbox | Named | Named | Named | Named | Named |
Example run, 6 of 50 prompts shown. Every answer is stored, so you can open the exact text behind each cell.
02Visibility index
Each part is measured per engine and averaged, so one busy engine can’t carry the score. Competitors get the same index, which sets your rank.
0.35V + 0.20S + 0.20P
+ 0.15C + 0.10T
Segments with fewer than 20 answers are marked low confidence, and pooled visibility comes with a 95% range.
03Engine scorecard
The five numbers agencies quote to clients, for every engine.
| Engine | Mention rate | Prompt coverage | Citation coverage | Share of voice | Accuracy |
|---|---|---|---|---|---|
| ChatGPT | 48% | 62% | 22% | 27% | 91% |
| Gemini | 36% | 46% | 14% | 21% | 84% |
| Perplexity | 55% | 70% | 38% | 30% | 95% |
| Claude | 42% | 54% | 18% | 24% | 88% |
| AI Overviews | 27% | 38% | 11% | 16% | 80% |
04Sources & accuracy
Every cited link is stored. Compare who the engines trust for your rivals, and check what they claim about you.
Sites cited when rivals are recommended, but never for you.
Claims about your prices, plans and features, checked against your site.
“Lumen Desk starts at $29 per seat and has no free trial.”
Your site says: $18 per seat, 14-day free trial.
Other wrong claims · accuracy 88%
05Action plan
Built from your own run: the questions you lose, the sites that skip you and the facts that are wrong. Each step says which part of the index it moves.
Adapts to your business
g2.com · capterra.com
Draft review request with AI06Progress & reports
Track daily, weekly or monthly. Progress compares like with like, and every run can become a report.
+6 this week
This week’s average against last week’s.
Example visibility index over time: 24, 27, 29, 30, 33, 35, 38, 41.
Pricing
Setup, the brand brief, prompt drafting and reports are free. You only pay for runs.
Buy any amount. Pay once.
$5per credit
$5010 runs
About 10 days of daily tracking for one brand, or 10 runs spread over several.
You pick the exact amount at checkout.
Daily tracking, renewed every month.
$139/ month
Billed monthly through Paddle.
Agency plans include runs too: Agency Starter 1 · Agency Pro 2 · Agency Scale 7 a month. New here? Create a free account first, then pay on the Billing page.
FAQ
New to GEO? Start with the overview. Need AI crawlers to reach your site first? See AI search readiness.
It asks every prompt in your approved set (50 by default) once on ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews with web search on, stores every answer, and scores it. One run uses 1 GEO credit.
Add your site and see every answer on your first run.