/how we do it
How we do it three movements · proprietary instruments

How we get you inside the answer.

Generative engines decide what to cite on three signals: can they extract you, do they trust the source, and is it recent enough. We don't guess at any of them. We measure, build, and sustain. Below is the full method, and the instruments we built to run it.

MOVEMENT_01

Diagnose. Measure where you actually stand.

Most agencies start by selling you content. We start by measuring you, across every public model, against the competitor you actually care about, on the queries that decide your category. You don't know what to fix until you know what models say.

What we measure

For each engagement, we run a baseline scan across ChatGPT, Claude, Perplexity, Gemini, Google AI, Copilot, Grok, and DeepSeek. Each model is queried with the prompts your buyers actually use, surfacing how often you appear, in what position, and with what sentiment relative to competitors.

  • visibility: how often you appear at all
  • position: where in the answer you land
  • sentiment: how you're characterized vs. rivals
  • sources: which citations the model leans on

Why it matters

The citation pool models draw from is narrower than the open web, and overlaps surprisingly little with traditional search rankings. Without a baseline you can't tell whether you're invisible, mis-cited, or losing share of voice to a single competitor.

28% of pages cited by ChatGPT have zero visibility in Google. Your SEO standing doesn't transfer automatically.

What you walk away with

A live dashboard of your position across all eight models, a competitor gap analysis, and a prioritized list of the queries we can move first.

baseline_scan.live
modelyour share
ChatGPT2 / 48
Claude3 / 48
Perplexity1 / 48
Gemini0 / 48
▸ Competitor A31 / 48
baseline · 48 buyer queries● live
MOVEMENT_02

Build. Make you extractable and citable.

Once we know the gap, we close it on two fronts at once: rebuild the site so models can cleanly extract what you offer, and manufacture the citable content (owned and third-party) that earns inclusion in answers.

Site structure: making you extractable

Generative engines prefer self-contained, semantically clean blocks with unambiguous claims, clear attributions, and structured data. Most marketing sites are written for humans and styled for designers. Neither serves a model trying to lift a single answer out of a page.

  • schema markup aligned to your business type
  • claim-and-evidence pairing on key pages
  • FAQ structure that maps to real buyer queries
  • internal linking that signals authority

Content: manufacturing the citations

Our Article Engine learns your brand voice, your industry, and your strengths, then writes current, fully-cited articles for first- and third-party publication. The goal isn't volume. It's publishing the artifacts that models reach for.

→ Recency and source diversity are dominant factors in citation selection, not raw domain authority.

What you walk away with

A rebuilt site that passes a GEO extraction audit, a content pipeline producing weekly cited articles, and an FAQ layer aligned to the buyer queries from Movement 01.

article_engine.live
outputstatus
"What is GEO and how it differs from SEO"published
"How AI assistants choose what to cite"published
▸ "Recency as a citation signal: 2026"drafting
"FAQ: GEO for restaurants"queued
"Outreach: third-party citation map"queued
brand voice · citation-checked● engine running
MOVEMENT_03

Sustain. Recency wins. Keep winning.

Citation isn't a one-shot prize. Models continuously re-rank what they trust, and freshness is one of the strongest signals. Sustain is the continuous loop: refreshing content, seeding third-party mentions, and watching the rankings shift daily.

Refresh: staying current

Content decays. Stats go stale, links break, claims age out. Our refresh loop scans your published library, flags decay, and updates the pieces that matter, keeping recency signals strong on the pages models already cite.

Third-party seeding

Models trust source diversity. We map the citation universe of your category (which publications, directories, and forums models actually pull from) and run targeted outreach to land mentions where it counts.

  • citation-universe mapping per vertical
  • targeted outreach to high-signal sources
  • guest-published cited articles

Daily tracking, daily decisions

Every morning your dashboard updates with overnight movement across all eight models. You see what shifted, why, and the next action, not a quarterly report card.

95% of pages cited in AI answers are less than 10 months old. Recency is not optional.
// instruments in this movement
daily_tracker.live
24h changedelta
ChatGPT: citation count+4
Claude: citation count+2
Perplexity: position↑ 03 → 01
▸ Share of answer14% → 22%
Gemini: pending refreshin queue
updated every 6h · 8 models● tracking
//The instruments

We didn't buy a platform. We built one.

Every movement above runs on instruments we built ourselves, because we understood the mechanism well enough to build them. A growing set, with more shipping.

[01]
article_engine

Learns your brand voice + industry, then writes current, fully-cited articles that weave in your strengths, for first- and third-party publication.

in production
[02]
sentiment_analysis

Queries the major models to surface what LLMs actually think, and say, about your brand.

in production
[03]
head_to_head

How often you appear versus a named competitor, plus the sentiment gap between you.

in production
[04]
visibility_analysis

How often you surface across the queries that define your industry: your true share of the answer.

in production
[05]
source_finder

Maps the citation universe of your category (which sources models trust), then targets them for mentions.

in production
[06]
site_restructuring

Rebuilds your site to best-practice GEO + SEO structure so models can extract and attribute it.

in production
[07]
content_refresh

Keeps your existing presence current and aligned. Recency is decisive for citation.

in production
[08]
faq_engine

Finds the questions buyers actually ask about you and your industry, answers them, publishes.

in production

// per-instrument detail pages shipping soon, more tools in development

// next

Want to see this method run on your brand?

Book a 30-min visibility assessment. We'll run a live baseline scan on your domain and walk through the first movement together, no pitch, just what we find.