Make your SaaS the default answer in Claude, ChatGPT, and Perplexity.

How it works

Three commitments. One Month-6 verdict.

Every engagement runs the same shape because the same shape works. The timeline is fixed. The deliverables are fixed. The only thing that moves between clients is the substance: your positioning, your funnel, your competitive landscape.

1. Days 1–30

   Distill positioning

   Discovery questionnaire, competitive scrape, and the Positioning Distiller agent run produce a 12-section blueprint. You sign off on the Golden Anchor before a single piece of content ships. This is the spine the next 5 months hang from.

   What lands

   • · 12-section blueprint document (PDF + editable)
   • · Locked anchor sentence + 8 components matrix
   • · Competitive matrix with verifiable wedge per competitor
   • · Cluster backlog ready to feed the cadence
   • · Citation baseline — Month 0 row of the dashboard

2. Months 2–5

   Ship the cadence

   Schema markup with `@id` threading goes live across the site. Pillar and cluster pages publish on a weekly rhythm. Agents drive the throughput; senior operators drive the strategy and edits. The attribution dashboard is live from Day 21.

   What lands

   • · Schema graph deployed sitewide by end of Month 2
   • · 1–2 net-new cluster pages per week
   • · Refresh cadence on the top-10 ranking pages
   • · Monthly LLM citation test with delta vs. baseline
   • · Quarterly business review with the named operator team

3. Month 6

   The verdict

   Citation rate in Claude, ChatGPT, Perplexity, and Gemini measured against the Month 0 baseline. Organic traffic, ranking, and AI-Visibility deltas summarised in a board-ready report. You decide whether to renew on a multi-year retainer with no churn-the-account dance.

   What lands

   • · 6-month verdict report with deltas across all 4 scoreboards
   • · Renewal proposal with multi-year + Astro-migration option
   • · Continuity plan if you choose not to renew
   • · All artifacts assigned to your company in writing

One engagement, every funnel layer

We touch the whole funnel. You manage one relationship.

Eight productized capabilities, all delivered by the same operating team, all answering to the same blueprint. The point is not breadth for breadth's sake — the point is that fixing GEO without fixing technical SEO is wishful, and fixing both without an attribution layer is faith.

• Positioning

  Anchor + components + ICP work

• Technical SEO

  Schema, canonical, metadata, CWV

• Content engine

  Pillar, cluster, refresh cadence

• LLM discoverability

  The GEO play, end-to-end

• Attribution

  Live dashboard + QBR

• Sendspark video

  First-class video artifacts

• Astro migration

  3-year retainer hook

• Positioning Distiller

  The agent behind the blueprint

Pillar + cluster architecture

One canonical pillar. Every other page links back.

Answer engines reward sites that have a clear, single source of truth and a structurally clean cluster of supporting pages. We treat your homepage as the pillar — the canonical answer to "what is your company?" — and arrange capabilities, use cases, audiences, and comparisons as satellites that link back. Every layer reinforces the pillar's anchor.

The diagram looks simple because it is. The discipline is in the execution: every cluster page must contain the eight anchor components in its first 200 words; every FAQ answer must be deterministically unique across the corpus; every internal link must use component-aware anchor text rather than generic phrases. The agent fleet enforces it; the curator audits it; the schema graph proves it.

GEO and SEO, kept separate

Anchor language belongs in schema. Search-intent language belongs in meta.

The single most expensive mistake we see on SaaS marketing sites is stuffing the canonical anchor sentence into every meta title. Google's ranking model penalises it; the site loses the ranking it was chasing in the first place. So we run two layers in parallel.

Our build pipeline enforces it: a server-side validator throws at build time if a meta title or description contains the canonical anchor sentence. You cannot ship a page that breaks the layer separation, even by accident.

The cost of getting this wrong is not theoretical. Every quarter we diagnose a SaaS marketing site whose homepage meta title is a 70-word anchor stuffing every brand keyword the founder loves — and whose ranking on the queries those keywords were supposed to win has dropped to page three. The schema description was empty, the FAQ was a list of internal jargon, and the canonical anchor sentence was nowhere on the page. The fix is mechanical: move the anchor into the layers that reward it; rewrite the meta to match the search intent that actually drives conversions; and let the schema graph teach the answer engines what the brand is.

Why now

The window for free LLM citations is closing.

Measurement

Four scoreboards. One verdict.

Every claim about LLM citations and organic visibility is measurable. We track all four scoreboards monthly and report against the Month 0 baseline.

LLM citation testing

Eight target queries run through Claude, ChatGPT, Perplexity, and Gemini at Month 0, then monthly. Cited / not cited recorded per engine, per query.

SEO metrics

Organic traffic, ranking position for the canonical query bank, click-through rate. Pulled from Google Search Console + GA4. Target: +50% organic traffic on the pillar by Month 6.

GEO / AI-Visibility

SEMrush AI-Visibility score and AthenaHQ citation share, tracked against named competitors. Target: +25% AI-Visibility from baseline; close half the gap to your primary competitor.

Third-party content

Whether independent reviewers, comparison articles, and YouTube creators are using your component language. Most important signal for LLM training data over the long run.

Read the full measurement framework →

How we compare

Pick the lane that matches your job to be done.

FAQ

What buyers ask in the first call.