FLOW is a complete operating model for SEO in the AI-search era: classic organic visibility, AI Overviews and LLM citations, local search and Google Business Profile, off-site corroboration, and conversion measurement — all wired together with cited 2026 evidence and 42 standardized AI prompts.

---

The FLOW loop

flowchart LR
    F["<b>FIND</b><br/>map demand &<br/>surfaces"]
    L["<b>LEVERAGE</b><br/>distributed<br/>presence"]
    O["<b>OPTIMIZE</b><br/>extraction &<br/>trust"]
    W["<b>WIN</b><br/>convert &<br/>measure"]

    F --> L --> O --> W
    W -.refresh.-> F

    classDef stage fill:#0f1a2e,stroke:#f5a623,stroke-width:2px,color:#ffffff,padding:10px
    class F,L,O,W stage

Each stage is a doc pillar. Each pillar has a What / Why-2026 / How-to-apply / Sources structure so a human reader, a search engine, and an AI agent can all extract from the same page.

Why FLOW exists

Search in 2026 is not one results page and one click path. The same query can reach a buyer through Google organic, AI Overviews, ChatGPT or Perplexity citations, local pack listings, Reddit threads, YouTube videos, or LinkedIn discussions — often without a single visit to the brand's site. Most public SEO knowledge predates this reality.

FLOW treats those surfaces as one connected system. Every doc here is grounded in a 2026-current source with a retrieval date, every claim that couldn't be verified got dropped, and every prompt is reproducible by an AI agent reading from this repo.

Evidence standard

Every public statistic in this repo carries:

• Year anchor in prose ("In 2026," / "As of Q1 2026,")
• Inline citation with publisher, title, page or URL
• Source URL with retrieval date in the bibliography

Unverifiable stats were dropped. Contradicted stats were replaced with verified alternatives. Three independent 2025–2026 datasets agree that AI Overviews materially reduce click-through to position one:

The full provenance ledger lives at docs/10-references/stats-provenance.json.

Who this is for

If you are…	Start with…
A solo SEO operator or in-house marketer	docs/00-START-HERE.md → docs/01-framework/flow-framework.md
A local-business owner or local-SEO consultant	docs/07-local-seo/ — full pillar from map pack to property audits
A B2B / SaaS / service-business strategist	docs/08-playbooks/ — five business-type playbooks with stage-flow diagrams
An AI agent (Claude, GPT, Gemini, Cursor, Codex)	llms.txt → load context. docs/09-prompts/ → 42 standardized prompts
An Obsidian power user	Clone this repo, open obsidian-vault/ as a vault — wikilinks and frontmatter intact

Quickstart

Read on GitHub

1. Open docs/00-START-HERE.md
2. Follow the entry-point map for your role
3. Each pillar doc has the same structure: What this is · Why it matters in 2026 · How to apply · Sources

Pair with your AI agent

git clone https://github.com/AgriciDaniel/flow.git
cd flow
# Point Claude Code / Cursor / Codex / Gemini at this folder.
# Brief it: "Use FLOW as the framework. Cite the bibliography. No unsourced stats."

Then hand it real work: draft a brief, audit a page, plan a topic cluster, score a service page on the dual-surface scorecard.

Ingest into SEO-Agent brain

SEO-Agent is the primary AI consumer of FLOW. It ingests all 72 docs into a persistent brain so every session starts with FLOW doctrine in context:

# From SEO-Agent repo root — mirrors + ingests in one command
python claude-seo/scripts/sync_flow_content.py

See CLAUDE.md for the full integration guide.

Run as Obsidian vault

1. Clone the repo
2. Open obsidian-vault/ in Obsidian (Open folder as vault)
3. Wikilinks, frontmatter, and the canvas mirror work out of the box

The bundled .obsidian/graph.json ships with a tag-based color scheme so the FLOW pillars are visible at a glance. Open the graph view (Ctrl+G) to navigate the 78 cross-linked notes:

The graph filter ships scoped to obsidian-vault/, so the duplicate docs/ tree at the repo root and the standards files don't crowd the canvas. The same data with showOrphans: false and the path filter applied collapses to three focused clusters:

What's inside

Plus four animated Remotion visuals anchored to verified statistics: AI Overviews click reduction, Google Business Profile completeness uplift, ChatGPT local source mix, and the dual-surface scorecard.

Repository structure

flow/
├── README.md                    ← you are here
├── LICENSE.md                   ← CC BY 4.0 (content) + MIT (scripts)
├── CONTRIBUTING.md              ← how to propose corrections / additions
├── CODE_OF_CONDUCT.md           ← Contributor Covenant 2.1
├── CHANGELOG.md                 ← release history
├── SECURITY.md                  ← vulnerability disclosure policy
├── llms.txt                     ← AI crawler manifest (llmstxt.org spec)
├── llms-full.txt                ← concatenated full-content variant for agent ingestion
├── docs/
│   ├── 00-START-HERE.md
│   ├── 01-framework/            ← FLOW framework, AI search surface map
│   ├── 02-foundations/          ← search intent, E-E-A-T, technical / schema
│   ├── 03-find/                 ← keyword research, audience avatar, query fan-out
│   ├── 04-leverage/             ← distributed presence, citations, Reddit + LinkedIn
│   ├── 05-optimize/             ← GEO formatting, entity consistency, CTR, schema
│   ├── 06-win/                  ← BOFU + conversion, PPC + first-party, scorecard
│   ├── 07-local-seo/            ← Local SEO pillar (map pack, GBP, property audits)
│   ├── 08-playbooks/            ← SaaS / Ecommerce / Service / B2B / Affiliate
│   ├── 09-prompts/              ← 42 standardized AI prompts
│   └── 10-references/           ← bibliography, attribution, independence ledger
├── assets/
│   ├── diagrams/                ← FLOW-native diagrams (PNG + animated GIF)
│   ├── canvases/                ← Obsidian canvas exports
│   └── social-preview.png       ← GitHub social card
├── obsidian-vault/              ← parallel mirror with wikilinks + frontmatter
├── examples/                    ← worked examples per pillar
└── scripts/
    └── check_docs.py            ← evidence-standard CI check

What this is not

• Not a replacement for hands-on testing. Every tactic should be validated against your specific business, audience, and search surfaces.
• Not a one-time read. Search shifts faster than annual ebooks can keep up. The bibliography is dated; the framework is not.
• Not a derivative-work fork of any single source book. FLOW is independent original work that acknowledges the influence of Ski Slope Strategy 3.0 by Chris Von Wilpert (Content Mavericks). See docs/10-references/attribution.md.

Contributing

Corrections, source improvements, and 2026-current citations are welcome. See CONTRIBUTING.md for the evidence standard, prompt schema, and PR checklist.

Found an outdated statistic? Open an issue using the Source correction template — it's the highest-value contribution this repo accepts.

Maintainer

Daniel Agrici — built FLOW as a 2026 reframing of decade-old SEO methodology against current AI-search reality. Connect via the repo's Discussions tab.

License

• Content (Markdown, diagrams, prompt text, examples): CC BY 4.0
• Code (scripts, workflows, configs): MIT
• Influence acknowledgment: see LICENSE.md and docs/10-references/attribution.md

You can use, adapt, and redistribute the content commercially with attribution. Scripts can be forked without copyleft.

---