The GEO Workbook: 30-Day AI Visibility Action Plan — 2026 Edition

The GEO Workbook is a free 30-day structured action plan for implementing Generative Engine Optimisation — one task per day, across four themed weeks, with fill-in templates, checklists, and tracking sheets for every stage. By Day 30, your site will have completed a full AI visibility audit, rewritten five pages for AI citation, implemented schema markup, built author identity signals, started off-page citation building, and set up ongoing monitoring. It is Book #4 in the GEO Lab Library.

The other GEO Lab books teach you what GEO is. This one makes you do it. No theory, no optional reading — just one task per day, a template to complete, and a measurable outcome for each of the 30 days.

Why a 30-Day Programme

GEO implementation fails for one reason more than any other: people understand the framework and then make no changes to their site. This is not a knowledge problem. Reading the GEO Stack page takes ten minutes. Understanding that AI engines cite extractable, entity-anchored, structurally clear content is not a complicated idea. The gap is between understanding and doing — and it is enormous.

The workbook closes that gap by removing all the decisions that cause paralysis. You do not decide what to audit on Day 1. You are told exactly what to audit, given a template to fill in, and told to stop when the template is complete. The same applies to every day across the 30. The decisions were made in the design of the programme. Your job is to show up and do the task.

Thirty days is long enough to make meaningful changes to a real site and short enough that the end is always visible. Week 2 is the most intensive — five pages rewritten from scratch using the GEO Writing Formula. That is the work most practitioners avoid indefinitely. In the workbook, it is just days 8 through 14.

How the Workbook Maps to the GEO Stack

The GEO Stack is a five-layer model for AI citation: Retrieval Probability, Extractability, Entity Reinforcement, Structural Authority, and System Memory, plus a Layer 0 covering infrastructure accessibility. The workbook does not teach the stack — it operationalises it. Each week targets a specific set of layers.

Week 1 (Audit) maps to Layer 0 and Layer 1. Before anything else, you need to know whether AI engines can access your pages at all, and whether your content currently appears in their retrieval sets for queries your audience uses. The technical health check on Day 4 covers Layer 0 directly: crawl access, page speed, robots.txt configuration, and structured data infrastructure. The AI visibility audit on Day 1 measures your current Layer 1 position — how often you appear versus competitors when AI engines retrieve sources for your target queries.

Week 2 (Rewrite) is Layer 2 — Extractability. This is the layer where most sites have the most room to improve. GEO Lab experiment data on extractability shows that the structural features of a passage — definition-first ordering, explicit labelling, short declarative sentences — have a measurable effect on whether AI engines pull from that passage or skip it. The GEO Writing Formula taught on Day 8 codifies those features into a rewrite process. Days 9 through 14 apply it to five real pages on your site.

Week 3 (Authority) maps to Layer 3 and Layer 4 — Entity Reinforcement and Structural Authority. Author schema, About page optimisation, and external citation building all feed the entity graph that AI engines use to assess source credibility. The internal linking audit on Day 22 addresses hub-and-spoke structure, which affects how citation authority distributes across a site.

Week 4 (Expand and Monitor) addresses Layer 5 — System Memory — and begins the measurement loop. Setting up citation monitoring across ChatGPT, Perplexity, and Gemini is not optional maintenance. It is the feedback mechanism that tells you whether the preceding 21 days of work moved the needle. The GEO Score assessment on Day 30 compares against the Day 0 baseline.

What the Data Behind This Programme Shows

The workbook was designed against a body of controlled experiments, not a theoretical framework. Several findings shaped the structure directly.

The most important is the two-stage citation funnel. Being retrieved is not the same as being cited. GEO Lab experiments consistently show that pages can appear in an AI engine’s retrieval set and still not be cited in the generated response — because the content is not extractable. A page can rank well in traditional search and be invisible in AI-generated answers. Week 2 of the workbook addresses this gap explicitly.

E027 found that Perplexity’s citation behaviour on proprietary-concept queries is deterministic: the same sources are cited across repeated queries over a 14-day window, with zero variance in the citation set. This has a practical implication. If Perplexity is not citing you for a query you should own, it will continue not citing you until something changes in your content or your entity signals — not until you wait longer or publish more. The workbook’s Day 1 audit is designed to identify those queries precisely so the Week 2 rewrites target them.

E042 found that ChatGPT’s citation gate is not binary. A minimum content length, named-platform context, and comparative framing together bypassed the gate that shorter, single-platform content could not. Week 2’s GEO Writing Formula incorporates those structural features — not because they guarantee citation, but because they remove the structural barriers that prevent it.

The baseline citation rate for commodity content across AI engines is effectively zero. Proprietary, non-commodity content — content that answers a specific question with a specific mechanism that no other page answers in the same way — earns citation rates above 80% on Perplexity T1 queries in GEO Lab measurements. The workbook’s approach to page selection (Days 9–14) is built around this distinction. You are not rewriting your five most-trafficked pages. You are rewriting your five most defensible pages.

What’s Inside the GEO Workbook

Who This Is For

The workbook assumes you have a published website, basic familiarity with your content management system, and access to at least one AI search tool. It does not assume technical skills — every schema task uses plugin-based implementation, covered in the companion GEO for WordPress guide. It does not assume any prior knowledge of GEO — the GEO Glossary covers every term used across the programme.

It is most useful for practitioners who have read about GEO and not yet acted on it, and for site owners who want a structured implementation programme rather than a list of recommendations to work through in no particular order. It is not a substitute for running your own experiments — that comes after Day 30, via GEO Experiments.

The workbook is free. No email required. No paid tools needed at any stage. The monitoring setup in Week 4 uses the free tiers of ChatGPT, Perplexity, and Gemini.

Frequently Asked Questions