Locked prompt set
Forty buyer questions across five types, versioned and frozen, so every month is measured against the same yardstick.
73% of cybersecurity vendors get zero ChatGPT citations in their own category
Security buyers ask AI first. Make sure it names you.
We measure how ChatGPT, Claude, Gemini, Google's AI and Perplexity answer your buyers' questions, show you who gets cited instead of you, and close the gap.
Built for Series A to B cybersecurity vendors
Citation gap mapQuery builderExport CSVEngine filter
| Buyer question | ChatGPT | Perplexity | Claude | Gemini | AI Overviews | AI Mode | Status |
|---|---|---|---|---|---|---|---|
| best MDR for a 300-person company without a SOC | Comp | Comp | — | Comp | Comp | Comp | Gap |
| CrowdStrike alternatives for a mid-market team | Comp | You | Comp | Comp | Comp | — | Partial |
| SIEM with SOC 2 evidence export | — | — | — | — | — | Comp | Open |
| how do I get 24/7 monitoring without an analyst | Comp | Comp | Comp | Comp | — | Comp | Gap |
| CNAPP that works with GKE and Azure | You | You | Comp | You | Comp | Comp | Partial |
| email security for a 40-person fintech | Comp | — | Comp | Comp | Comp | Comp | Gap |
40 prompts tracked · 5 engines · 5 runs each · refreshed monthly
What we do
Every layer of AI visibility, measured before it is managed.
A locked question set, honest rates instead of screenshots, and a plan mapped to the pages the engines actually cite.
Five engines, many runs
Each question is asked repeatedly on ChatGPT, Claude, Gemini, Google's AI surfaces and Perplexity. You get rates with intervals, never a single screenshot.
ChatGPT11/40
Perplexity14/40
Claude4/40
Gemini9/40
AI Overviews6/40
AI Mode5/40
Source gap map
Every page the engines cite for your category, grouped by who owns it. Most of the time it is review sites, forums and a competitor's comparison page, not you.
reddit.com/r/cybersecurityReddit19 answers
g2.com/categories/mdrReview site16 answers
competitor-a.com/vs/youCompetitor12 answers
esecurityplanet.com/best-mdrEditorial9 answers
attack.mitre.orgStandards6 answers
Competitor teardown
What the most-cited competitor publishes, where it gets placed, and which of those placements are open to you.
Competitor A31/40
Competitor B22/40
You14/40
Competitor C9/40
Attribution kit
A GA4 channel for AI referrals, a crawler-log parser that proves ingestion, a self-reported field, and a branded-search baseline installed before work starts.
GA4 channelAI Assistant
Crawler log parserGPTBot, ClaudeBot
Self-reported fieldHow did you hear
Branded searchbaseline captured
Your strategist
A human owns the plan: weekly updates, a monthly re-run of the locked set, and a KPI report that says what moved and why.
✓Weekly update sent
✓Locked set re-run
✓Delta report
✓Monthly KPI report
Why teams run on Prescaled
Rates, not screenshots
AI answers change between runs and between days. We report what a buyer is likely to see, with the uncertainty on the page.
Citation gap map
Citation gap mapQuery builderExport CSVEngine filter
| Buyer question | ChatGPT | Perplexity | Claude | Gemini | AI Overviews | AI Mode | Status |
|---|---|---|---|---|---|---|---|
| best MDR for a 300-person company without a SOC | Comp | Comp | — | Comp | Comp | Comp | Gap |
| CrowdStrike alternatives for a mid-market team | Comp | You | Comp | Comp | Comp | — | Partial |
| SIEM with SOC 2 evidence export | — | — | — | — | — | Comp | Open |
| how do I get 24/7 monitoring without an analyst | Comp | Comp | Comp | Comp | — | Comp | Gap |
| CNAPP that works with GKE and Azure | You | You | Comp | You | Comp | Comp | Partial |
| email security for a 40-person fintech | Comp | — | Comp | Comp | Comp | Comp | Gap |
40 prompts tracked · 5 engines · 5 runs each · refreshed monthly
Visibility report
Visibility report · mention rate across engines
You14/40
Competitor A31/40
Competitor B22/40
Competitor C9/40
Source gap map
Source gap map · pages cited where you are absent
| Page | Type | Cited in |
|---|---|---|
| reddit.com/r/cybersecurity/... | 19 answers | |
| g2.com/categories/mdr | Review site | 16 answers |
| competitor-a.com/vs/you | Competitor | 12 answers |
| esecurityplanet.com/best-mdr | Editorial | 9 answers |
| attack.mitre.org/... | Standards | 6 answers |
Monthly delta
Monthly delta · baseline vs month 2
| Engine | Baseline | Month 2 | Change |
|---|---|---|---|
| ChatGPT | 5/40 | 11/40 | +6 |
| Perplexity | 9/40 | 14/40 | +5 |
| Claude | 3/40 | 4/40 | +1 |
| AI Overviews | 2/23 | 6/25 | +4 |
Model change noted: ChatGPT proxy moved to a new snapshot on the 14th.
Questions measured the way buyers ask themCompetitors cited, page by page, engine by engineA plan for every gap, tracked monthly
What the data says
The answers are already being given. Mostly by other people's pages.
73%
of cybersecurity vendors get zero ChatGPT citations in their category
GrackerAI benchmark, 100 vendors, 250 prompts, 6 engines, 20263 to 12%
of citations on non-branded prompts point at the brand's own site
Ranqo (2.9%, 150k citations) and Derivatex (11.6%), 2026FAQ
Questions we get on every first call.
General
What is Prescaled?
An AI-visibility agency for cybersecurity vendors. We measure how AI assistants answer your buyers' questions, show which competitors and sources get cited instead of you, and run the work that closes the gap. Software does the measuring; a human owns the plan.
How is this different from SEO?
SEO is ranking on Google's blue links. AI visibility is being named and cited when a buyer asks ChatGPT, Claude, Gemini, Google's AI or Perplexity the same question. The two overlap, but the scoreboards are separate, and AI answers lean on third-party pages far more than on your own site.
Who is this for?
Series A to B cybersecurity vendors with a marketing lead and a real category: MDR, SIEM, CNAPP, compliance automation, email security, identity, and the like. If your buyers evaluate vendors with an AI assistant, the questions we track are the ones they ask.
How it works
How do you measure visibility?
We write 40 buyer questions with you, lock them, and ask each one five times on five engines from a clean session with no memory or history. A brand counts only when the answer names it. Every figure is a rate over runs with a 95% interval.
Why rates instead of screenshots?
The same question gets different answers minute to minute and day to day; roughly two thirds of cited sources change between days. One screenshot proves nothing. Rates with intervals tell you what a buyer is likely to see.
Which engines do you cover?
ChatGPT and Claude through their APIs with web search, Perplexity through its API, Gemini through Google's API with search grounding, and Google AI Overviews and AI Mode captured from live US search results. Model versions are printed in every report, along with the caveat that API results approximate the consumer apps.
What do you not promise?
A revenue figure from AI referrals, placement in any specific answer, clean last-click attribution, or results by a date. Engines change and most AI-influenced visits arrive as direct traffic. We promise honest measurement on a locked set, improvement over the term, accurate representation, and growth in the third-party pages that mention you.
Reporting
What do we get each month?
A re-run of the locked set with a delta report (what moved, per engine, with model-change notes), an updated source gap map, the placements and content shipped against it, and a weekly update from your strategist.
How do you attribute results?
Three layers, never summed into one number: verified AI referrals in your analytics, an estimated share from AI-session ratios, and influence read from branded-search and direct lift. The instruments are installed before the work starts so there is a baseline.
Find out where you're invisible.
We'll map the questions your buyers ask, show you who gets cited today, and tell you which gaps are winnable.