Intermediate · 7h 26m · 32 lessons
GEO — Generative Engine Optimisation
Getting cited when a machine answers instead of listing.
What you'll learn
- ✦Explain why retrieval competes on passages rather than pages, and what follows from that
- ✦Write chunks that answer directly, stand alone, and carry a specific figure
- ✦Tell parametric knowledge from retrieved knowledge, and know which one you can influence
- ✦Audit AI crawler access separately from search crawler access
- ✦Build a consistent entity footprint across sources you don't control
- ✦Diagnose and fix an answer that describes your product incorrectly
- ✦Plan content as question clusters so you stay present through a conversation
- ✦Measure with a question panel, and know what size of change yours can actually detect
- ✦Recognise the standard GEO sales pitches and why each one is unfalsifiable
Part of these career paths
- Search Specialist path · course 2 of 5
Certification projects
Build 6 projects to earn your certificate
Hands-on work that proves you can apply what you learned — part of the certificate requirements.
Syllabus
Module 1 · 58m
What answer engines do Free
What you'll be able to do
- Explain why retrieval scores individual passages rather than whole pages, and why that makes citation more winner-take-all than search ranking.
- Tell parametric knowledge (fixed at training) apart from retrieved knowledge (fetched now), and know which one you can actually influence.
- Judge whether GEO is working by segmenting assistant-referred traffic and its conversion rate, not by session count alone.
- Rewrite an opening passage to answer directly, in a self-contained way, with a specific figure.
Module 2 · 55m
Retrieval and access Locked
What you'll be able to do
- Explain why a page can hold the right answer and still fail to be cited, and rewrite a paragraph so its chunk stands alone.
- Audit a robots.txt file to see which AI crawlers are blocked, and separate a training-crawler decision from a retrieval-fetcher decision.
- Decide which pages actually need a freshness update, based on whether their answer changed rather than their age.
- Diagnose why an already-cited passage still ranks low, and raise its rank by adding a direct answer and a figure.
Module 3 · 57m
Entities and structure Locked
What you'll be able to do
- Explain the difference between a name (a string) and an entity, and what makes an entity resolvable.
- Explain what structured data does and does not do for GEO, and mark up an Organization and its key facts.
- Audit every place a company is described, find the contradictions, and write one consistent description to use everywhere.
- Rewrite brand copy so it states category, audience, price and named alternatives, to compete for a recommendation question.
Module 4 · 56m
Being the source Locked
What you'll be able to do
- Explain why original data does not automatically score higher in retrieval, and rewrite a finding to answer first and cite itself second.
- Judge which credibility signals (named authors, your own numbers, stated limitations) are hardest to fake, and use them in advisory content.
- Rewrite a hedged or compound claim into a quotable, self-contained sentence with a value in it.
- Diagnose why a high-authority official source can still lose to a more direct passage, and add original data without losing rank.
Module 5 · 56m
Brand presence in answers Locked
What you'll be able to do
- Judge whether a brand's problem is being described as nothing or being described wrongly, and pick the matching fix.
- Diagnose a wrong AI answer about a company by checking its cited source first, before assuming the model invented it.
- Explain why most citations about a company come from sources it doesn't own, and claim or correct the ones that matter most.
- Audit how an assistant answers the five questions a buyer actually asks about a product, and fix the page that should answer each one.
Module 6 · 56m
The conversation, not the query Locked
What you'll be able to do
- Explain why the requirements shared across answer surfaces matter more than any one surface's current quirks.
- Explain why winning the first question in a conversation does not win the follow-up questions, and plan for question clusters.
- Compare a broad, shallow content strategy against a narrow, deep one, and explain why depth wins more of a conversation.
- Extend an existing page to answer a follow-up question without rewriting what already works.
Module 7 · 54m
Measuring GEO Locked
What you'll be able to do
- Explain why GEO cannot be rank-tracked like search, and build a fixed question panel instead of trusting a single observation.
- Calculate the smallest change a question panel of a given size can actually detect, and report a rate with its interval.
- Build an evidence stack (referrals, citation rate, branded search, self-reported attribution) for a channel that hides most of its effect.
- Build a measurement plan with the fields to record per question, take the baseline, and set a review cadence.
Module 8 · 54m
Strategy: GEO and SEO together Locked
What you'll be able to do
- List the six things GEO adds on top of good SEO, and explain why none of them conflicts with search performance.
- Recognise the common GEO sales pitches, and apply one test ("how would I know if you were wrong, and whose measurement is it?") to each.
- Sequence a 90-day GEO plan by speed of return, and explain why AI crawler access is a gate rather than a lever.
- Tell a durable GEO principle from a configuration detail that's likely to change, and know what to recheck and how often.