Vertical Insights · Founding access
The AEO research for your category, as we discover it
Every week, Quoted measures what content features actually keep pages cited by AI search engines — and which ones don't survive scrutiny. Vertical Insights gives you read-only access to that research for a single category, updated as the data lands.
Join the founding waitlist →Early access to a live research program. Every finding ships with its evidence tier on the label, so you always know how hard you can lean on it.
Two categories are live today: cybersecurity and martech
Cybersecurity
Citations tracked weekly across ChatGPT search, Claude search, and Perplexity, on a CISO-grade prompt set.
Martech
The same study on a martech prompt set. Expanding toward broader B2B SaaS coverage as the corpus grows.
You're subscribing to the research, not a monitoring tool
Not a monitoring tool for your URLs — the research itself. We run the same longitudinal study across a vertical's content and report which features track with citations that persist, graded by how much evidence stands behind each one. You get the WHY behind staying cited, scoped to the category your buyers actually search in.
Want a vertical that isn't here yet? Tell us on the waitlist — we open new verticals based on where the demand is.
Most AEO tools tell you if you're cited. We tell you why citations last.
Most AEO platforms answer a binary question: are you being cited? That's useful, but it's a snapshot. It tells you what happened, not why—and not what to do about it.
Our research is built around a different question: what content characteristics predict whether a citation persists or decays over time? Answering that requires longitudinal data, structural analysis, and statistical rigor—not a single crawl.
We test the 100+ questions your buyers actually ask AI
Every research program starts with a prompt set—typically 100+ questions per vertical, covering the full buyer journey:
"What is [concept]?"
Questions from buyers just entering the market. These prompts test whether AI engines recognize a brand as a foundational source in the space.
"How does X differ from Y?"
Questions from buyers building a shortlist. These prompts reveal whether AI engines trust one explanation over a competitor's content.
"X vs. Y" and "best [category]"
Prompts from buyers evaluating options. High commercial intent. These are the citations that directly influence purchase decisions.
"Is [product] worth it for [use case]?"
Prompts from buyers ready to commit. Being cited here means being part of the final consideration set—the last mile before purchase.
We run every prompt three ways—because a buyer’s real question and a generic one share almost none of the same citations.
How a buyer actually phrases the question matters as much as the question itself. We tested this directly. The result was sharp enough that it changed how we run every engagement.
In a 45-call smoke test across cybersecurity questions, we ran each prompt three ways: generic (the bare question), persona-prefixed (with a one-line buyer-role prefix), and persona-rewritten (the question phrased the way a real healthcare-CISO would type it). On average, generic and persona-rewritten variants of the same underlying question shared 0% of their citations. Two parallel universes of sources for the same buyer intent.
This finding is why every prompt in a vertical’s Index is tracked in all three variants every week. Generic-prompt monitoring (what Profound, Stacker, and BrightEdge ship by default) measures a different citation graph than the one buyers actually encounter in their own sessions.
Read the full methodology and raw data: Smoke Test: Is Your AEO Software Tracking Real Citations? →
What this means for the findings: Every result is reported across all three variant types per prompt, not pooled—so you can see which content patterns hold in the generic citation universe AND in the persona-aware one buyers actually encounter, and where the two diverge.
How we turn weeks of citation data into content rules you can act on
Raw citation data tells you what happened. Our analysis layer tells you why. We call the output citation contributors—the specific content attributes that statistically correlate with citation persistence in a given vertical.
Feature extraction
Every cited URL is analyzed across 50+ structural and contextual features. Some are obvious (word count, schema markup). Others are specific to how AI engines process content and are part of our proprietary research.
Correlation analysis
We compare the features of URLs that persist in citations against those that decay. Over weeks and months of longitudinal data, statistical patterns emerge—certain content characteristics consistently predict persistence, and they vary by vertical.
Playbook generation
The citation contributors become prescriptive guidance: what to publish, how to structure it, which gaps to fill. These aren't generic best practices—they're specific to the vertical, graded by evidence, and drawn from weeks of longitudinal data.
What you get in every weekly update
Graded findings
Each pattern is tiered by evidence—from “recommend” down to “worth a test”—so you never act on a coin flip.
The reasoning, not just the result
Effect sizes, sample sizes, the engines it held on, and the confound checks it survived.
What we threw out
The findings that looked strong and failed scrutiny—so you don’t chase them either.
A working playbook
The findings translated into concrete, do-this-Monday content guidance for the vertical.
Founding access, and when you actually pay
Get in before the paywall. Founding members lock first access to every vertical we open — and pay nothing until the research is worth paying for.
- First access the moment your vertical's findings clear the confidence bar.
- The methodology, raw data, and caveats behind every finding — not just the takeaway.
- A direct line to shape which verticals we open next.
- No charge until the research is worth charging for.
Join the waitlist for first access
One category, read-only access to the research. We'll email you up to four times a month when findings are ready.