Methodology

3 phases, 9 steps:
how we make brands visible in AI search.

Our pipeline is not a black box. It is documented and structured so that every step is traceable on its own. We invest real audit hours before we make an offer — and we close this complete audit workshop free of charge.

  • 3 phases × 3 steps (canonical)
  • 1.5-hour workshop · digital · screen-shared
  • Customer-owned token
  • Concrete offer after the workshop
Pipeline visual

Three phases. Three steps. One basis for decisions.

The pipeline runs strictly in sequence. Every phase delivers an output that the next phase needs as input. Phase 1 builds context, Phase 2 understands the market, Phase 3 produces the outputs you and your agency can actually work with.

Phase 01

Build context

We build the knowledge graph of your brand from publicly available sources.

  1. 01 Website audit
  2. 02 Build Knowledge Graph
  3. 03 Aggregate reviews
Phase 02

Understand the market

We measure how generative engines currently talk about you and your market.

  1. 04 Condense profile
  2. 05 Research prompts
  3. 06 Competitive analysis
Phase 03

Generate outputs

We condense the full picture into concrete, customer-owned outputs.

  1. 07 Condense
  2. 08 Personas
  3. 09 FAQs & roadmap
In detail

What we do at every step, with which tools, with which output.

Each of the nine steps has a clearly defined input, a concrete tool selection and a measurable output. At the end of the pipeline, you do not see a generic presentation, but your brand in the industry-specific knowledge graph — condensed into personas, FAQs, roadmap and action tickets.

01 Phase 01 · Build context

Website audit

We crawl your domain, read Schema.org markup, robots.txt, llms.txt and meta tags. We measure which content is machine-readable for generative engines and which mandatory codes are missing.

Tools
Web crawler · Schema.org validator · llms.txt check
Output
Inventory of all mandatory codes (JSON-LD, llms.txt, robots.txt, FAQPage) — found / missing.
02 Phase 01 · Build context

Build Knowledge Graph

We anchor your brand as an entity in the industry-specific knowledge graph. Clusters, node types, disambiguators and sources are modeled along your industry schema.

Tools
Industry schema (KG library) · Perplexity AI · Vertex AI
Output
Complete KG dataset: sub-clusters, entities, relationships — as visualization, JSON-LD and plain text.
03 Phase 01 · Build context

Aggregate reviews

We aggregate public reviews from the platforms relevant to your industry, reading sentiment, topic clusters and the consistency profile across sources.

Tools
Google · TripAdvisor · Booking · HolidayCheck · Trustpilot · Google Business Profile
Output
Sentiment profile per platform and across sources — including topic tags and contradictions.
04 Phase 02 · Understand the market

Condense profile

We consolidate the outside-in picture of your brand from website, KG and reviews into a profile statement. We mark self-image vs. outside-image — where does it contradict?

Tools
GPT-4o · Perplexity · Sistrix
Output
Profile statement with self/outside comparison, specialization clarity and brand-coherence score.
05 Phase 02 · Understand the market

Research prompts

We craft 20–50 realistic research prompts that typical users in your industry would actually ask AI — categorized as brand, brand-mention and non-brand.

Tools
Google Keyword Planner · Sistrix · Perplexity topic suggestions
Output
Prompt library 20–50 queries, classified as brand · brand-mention · non-brand.
06 Phase 02 · Understand the market

Competitive analysis

We send the research prompts to ChatGPT, Perplexity and Google AI Mode. In every answer we mark: where are you mentioned? Who are the competitors? Which sources are cited?

Tools
ChatGPT · Perplexity · Google AI Mode · Helicone logging
Output
AI-visibility audit: Mention Rate, Citation Rate, Ø position, Share of Voice — per query, per engine.
07 Phase 03 · Generate outputs

Condense

We condense all findings into a composite GEO Score and a SEO × GEO quadrant-position indicator. Where do you stand objectively?

Tools
Custom scoring engine · Sistrix visibility index · GEO composite
Output
GEO Score (40% AI visibility · 25% KG coverage · 15% E-E-A-T · 20% mandatory codes) and quadrant position.
08 Phase 03 · Generate outputs

Personas

We generate 3–6 personas per industry, validated against your profile and reviews. Each persona comes with a bio, a typical AI question, needs — as a customer-owned asset.

Tools
GPT-4o · validation loop against KG + reviews
Output
Persona set of 3–6 personas, each with bio, AI query, needs, frame tag. Importance slider in the workshop.
09 Phase 03 · Generate outputs

FAQs & roadmap

We generate the FAQ library (brand · brand-mention · non-brand), a 90-day roadmap with quick wins and strategic actions, plus concrete code optimizations (Schema.org, meta, content).

Tools
GPT-4o · Helicone · custom roadmap generator
Output
FAQ library + 90-day roadmap + action tickets P1/P2 + code snippets for your site.
USP

„We invest real audit hours before every sale — free of charge. No mini audit, no self-serve teaser. Full pipeline, complete outputs, customer-owned token. Concrete offer after the workshop, project-specific."

  • Full Honest first analysis — no artificial limits
  • Pilot Concrete modeling offer directly from the specialist
  • 48 h Response time after audit request — human, not bot
  • 1.5 h Workshop digital, screen-shared, no slide deck
Workshop

1.5 hours, five blocks, zero slide deck.

The workshop is not a presentation — it is a joint walkthrough of your finished analysis. Concepts are explained exactly where they become relevant. We validate personas and FAQs together — you correct live, we capture live.

  1. 01
    Block 01 · ~10 min
    Foundations & the language of AI
    Why GEO, why now, what is a Knowledge Graph, what is an FAQ schema. Live on your data, not abstract.
  2. 02
    Block 02 · ~10 min
    Our methodology
    3 phases × 3 steps. Slide 13 as anchor visual. What you already have, what comes next.
  3. 03
    Block 03 · ~25 min
    Your project live
    Scores, quadrant, research prompts, AI visibility, gaps. We navigate through your customer portal — you see your brand inside the industry knowledge graph.
  4. 04
    Block 04 · ~30 min · Core
    Validate personas together
    The most important block. Importance slider, accept or reject per persona, notes live. Personas are yours. You correct, we learn alongside.
  5. 05
    Block 05 · ~15 min
    What we do next
    Persona review, offer, implementation, optional retainer. Clear next steps with time expectations.
Customer phases

What happens when — three clearly separated stages.

After the workshop, your customer token persists. What happens after Phase 1 is up to you. We give you a clear recommendation — not a pressure-sales pipeline.

Phase 1 Live now · token active

You validate live.

  • Validate personas with importance slider, accept or reject, note
  • Explore FAQ samples with filter, accept or reject, note
  • Customer token stays persistent, shareable, bookmarkable

Note: FAQs still rely on un-approved personas — they are fully reworked once personas are signed off.

Phase 2 After the meeting

Tailored offer.

  • Offer based on your roadmap actions — industry- and size-specific
  • Quick wins first, strategic actions staged by impact × effort
  • Setup fee + monthly platform subscription — itemized separately

Modeling investment is concretized here — not pre-flat-rated.

Phase 3 After commission · 5 building blocks

2 mandatory + 3 optional.

  • 01 · Mandatory AI conception & data model Knowledge Graph + personas + FAQ/prompt model
  • 02 · Mandatory Integration into your website Schema code + FAQ module
  • 03 · Optional FAQ module Accordion display of frequent questions
  • 04 · Optional Platform subscription
  • 05 · Optional Agency retainer SEO/GEO/content/AI coaching by your agency
Investment architecture

Transparent, staged, freely combinable.

Layer 0 is always free — you pay nothing for audit and workshop. Layer 1 (modeling + integration) is quoted in the post-workshop offer, industry- and size-specific. Layer 2 are the ongoing services you combine freely — nothing is bundled.

Layer 0

Audit + workshop

0

Free of charge. Full pipeline. 1.5-hour workshop. Customer token.

Layer 1

Modeling + integration

Industry-fair

Mandatory blocks 01 + 02. One-time. Concretized in the post-workshop offer.

Layer 2a · Platform

Maintenance · Monitoring · Enterprise

  • 69 €/month · Maintenance · Auto-updates KG/FAQs/Schema
  • 149 €/month · Monitoring · + AI visibility tracking + competitor alerts
  • 199 €/month · Enterprise · + SLA + custom endpoints + multi-property

Freely selectable. Contract directly with GEOLYZER.

Layer 2b · Optional

Agency retainer

0.5–2.5 k €/month

SEO/GEO/content/AI coaching by your agency — fully optional.

Note Layer 1 is intentionally not flat-rated — modeling scales with industry complexity and property count. Concrete numbers only after the workshop.

Next step

See our methodology live on your own brand.

We start with a free audit for your industry. 48-hour response time, 1.5-hour workshop, customer token stays with you — whether you commission later or not.