web analytics
Framework · Substance first · Production-tested

AEO/LLMO Content Optimisation Framework.

A practitioner’s framework for structuring B2B content so search engines and AI systems can parse, extract, and cite it. Substance first, structure second, and a way to measure whether any of it actually worked.

Built by Solange Rainha

The first version of this framework leaned hard on extractability mechanics. This version reorders it around what the research, and my own results, say earns a citation: the substance comes first, the structure delivers it, and citation tracking tells you whether any of it worked. It’s based on real production workflows, not theory.

The fundamentals

Two disciplines, one framework.

AEO and LLMO solve the same core problem from different angles: making your content the answer, whether it’s served in a search snippet or cited by an AI model.

AEO

Answer Engine Optimisation

Optimising content for featured snippets, People Also Ask boxes, and zero-click search results. The goal: your content is the answer Google surfaces directly, without requiring a click-through. That takes explicit definitions, structured data, and content that meets search intent in parseable formats.

LLMO

Large Language Model Optimisation

Structuring content so AI systems (ChatGPT, Claude, Perplexity, Gemini) can parse, extract, and cite it when generating answers. The goal: your content becomes a source these models pull from and attribute. That takes citation-worthy claims, named sourcing, and substance a model can’t reconstruct without you.

The operating order

Substance, then structure, then measurement.

Most AEO programs run this backwards. They start with the formatting because it’s cheap and checkable, then wonder why the citations never come. The order below is the whole point of the framework, because the work that earns a citation is the work that’s easiest to skip.

01

Substance: what earns the citation

The claims, data, sources, and point of view that give an AI model a reason to cite you over ten interchangeable pages. This is the part teams skip, so it goes first.

02

Structure: how it gets extracted

The formatting that lets a machine lift your substance cleanly: front-loaded answers, question headings, self-contained sections. Necessary and fast, but second-order. It delivers the substance; it can’t replace it.

03

Measurement: whether it worked

Tracking actual AI citations, not audit scores. A green audit tells you the structure is tidy. It tells you nothing about whether you’re being cited, which is the only outcome that pays.

Layer 1 · The citation drivers

Substance pillars: give a model a reason.

These are the levers with the biggest payoff and the ones the popular playbook treats as optional. Every one requires that you’ve done the work: the research, a position you can defend, the credible voices in your space quoted accurately. None of it is schema or heading format.

01 · LLMO

Original argument & point of view

The piece says something a model can’t pull from the public soup of recycled takes: original observation, a position on what works for whom, something that came out of actual work. If the page only repeats what ten others already say, there’s no reason to cite this one.

02 · LLMO

Statistics & specific data

Concrete, attributed numbers move AI visibility more than any formatting change. Adding statistics improved citation likelihood by around 41% in the Princeton GEO study. Vague claims get skipped; specific ones get pulled into the answer.

03 · LLMO

Named sources & quotations

Cite credible third parties by name and quote recognised voices accurately. Citing external sources lifted visibility by up to 115% for lower-ranked content; quoting named sources by around 28%. Pointing at other people’s authority is how you build your own.

04 · AEO LLMO

Comparison tables & structured evidence

Content with comparison tables earns more citations than text-only equivalents. A table of approaches with their trade-offs hands a model something concrete to lift and attribute, and it reads as evidence rather than opinion.

05 · LLMO

Current, verifiable sourcing

Sources that are recent and check out. Citation rates shift month to month and engines change, so dated or unverifiable claims age out of the answer set fast. Recency is part of credibility.

The evidence

What the research says drives citation.

The most cited measurement of what moves AI citation is the Princeton GEO study (Aggarwal et al., presented at ACM KDD 2024), which tested content strategies across 10,000 queries on multiple generative engines. The strategies that won were substance-side, not structure-side. That finding is the spine of this framework’s ordering.

~41%

visibility lift from adding relevant statistics to content

Princeton GEO study, KDD 2024
~28%

lift from adding direct quotations from named sources

Princeton GEO study, KDD 2024
up to 115%

citation lift from citing external sources, for lower-ranked content

Princeton GEO study, KDD 2024
0 / worse

adding more words did nothing; keyword stuffing scored below the baseline

Princeton GEO study, KDD 2024
Layer 2 · The delivery layer

Structure pillars: get out of the way.

I build every one of these into every piece, so this isn’t a knock on structure. It’s about sequence. Structure is fast and mechanical, which is exactly why it’s second-order: it delivers the substance underneath it. A green checklist on thin content is a well-formatted way to stay invisible.

01 · AEO LLMO

Explicit definitions

Every key concept and term defined clearly at first use, written to stand alone if the sentence gets pulled out of context entirely. No assumed knowledge, no definition that only makes sense inside its paragraph.

02 · AEO LLMO

Extractable summaries

The first 150 words contain a self-contained answer to the core question, substance included. If a snippet or an AI response grabs only the opening, the reader still gets the answer instead of a tease.

03 · AEO

Question-format headings

H2s written as the questions people actually ask. This maps directly to People Also Ask queries and gives a model clear context about what each section answers.

04 · AEO LLMO

Structured FAQ sections

A self-contained FAQ block near the end, each answer concise and citable without surrounding context, formatted for schema markup and zero-click extraction.

05 · AEO LLMO

Cross-reference layer

Pieces reference each other instead of repeating themselves, building a linked knowledge graph that crawlers and models can traverse to assess your depth on a topic.

Applied structure

The content blueprint.

How the two layers translate into a repeatable structure applied to every piece. Same documented ruleset, reordered so the substance decisions come before the formatting ones.

01

Substance: claim and evidence first

Before any formatting, decide the specific attributable claims this piece makes that nobody else is making, the data it cites, and the named sources behind it.

02

Opening: extractable summary (first 150 words)

A self-contained answer to the core question, substance included, so a snippet or AI response that grabs only this block still gets the answer. No throat-clearing.

03

Body: question-format H2 sections

Each section answers a query people actually type, definitions front-loaded, every claim standalone and attributed to its source within the content itself.

04

FAQ block: three self-contained Q&A pairs

Concise, citable out of context, formatted for schema markup and PAA extraction. Each answer works if it’s the only thing a model reads.

05

Cross-reference layer

Link related pieces to each other to build a knowledge graph crawlers and models can traverse for topical authority. Reference instead of repeating.

06

Quality gate: verify, then track

Every claim traced to documentation, data, or stakeholder confirmation before publish. After publish, track actual AI citations rather than audit completeness.

Workflow

How it fits into content production.

The framework runs from the first step of the workflow, and it ends with the step most teams skip: checking whether you were actually cited.

1

Research & intent

Keyword and intent mapping, plus what AI models already say about the topic and where the gap is.

2

Substance

Decide the original claim, gather the data, line up named sources, build the comparison table. The part that earns the citation.

3

Structure

Front-load the answer, write headings as questions people actually ask, keep sections self-contained so they survive extraction.

4

Verify

Source log check: every claim traced to documentation, data, or stakeholder confirmation before it ships.

5

Optimise

Meta elements, schema markup, internal cross-references, FAQ formatting, readability balance.

6

Measure

Track whether AI systems actually cite the piece. That’s the scoreboard, not the audit score.

Made concrete

What bad AEO looks like versus good.

“Add substance” is true and useless at the same time, so here are two versions of the same article. Both are well-formatted. One of them said something.

Bad AEO: structure on top of nothing

  • Generic claims any competitor’s post also makes
  • No data, no named sources, nothing specific
  • Wrapped in FAQ schema and question-format headings
  • Scores 95/100 on the audit
  • Never cited, because it offers a model nothing it can’t get elsewhere

Good AEO: substance the structure delivers

  • A specific, attributed claim: named study, percentage, year
  • A recognised practitioner quoted by name
  • A comparison table with trade-offs
  • An argument about what works for which stage
  • Clean structure carrying all of it, so there’s a reason to cite this page
The shift

Traditional SEO vs. AEO/LLMO content.

The difference is structural: content restructured so machines can extract value without a human reading the full page, and given something worth extracting once they do.

Traditional SEO content

  • Keyword density as the primary signal
  • Value depends on the click-through
  • Definitions buried in context
  • Stats with no visible sources
  • Generic headings (“Benefits of X”)
  • Content that exists in isolation

AEO/LLMO-optimised content

  • Citability as the primary signal
  • Zero-click value through direct answers
  • Definitions front-loaded and standalone
  • Every claim attributed and verifiable
  • Question-format headings matching queries people type
  • A cross-referenced knowledge graph
Why this matters

The ground has already shifted.

Discovery is moving off the blue link and into answers, which is why citability now matters as much as ranking.

58.5%

of US Google searches end without a click

Semrush, 2025
73%

of B2B buyers use AI tools in purchase research

Averi, March 2026
25%

predicted drop in traditional search volume by 2026

Gartner, 2024
Where teams go wrong

Six mistakes that kill AI discoverability.

Most teams approach AEO as an add-on to their existing SEO workflow, in the wrong order. That’s the first problem. Here are the rest.

01

Doing the structure first

The intuitive order is to write the content, then bolt on schema, question headings, and an FAQ. That sequence optimises the cheapest, most measurable layer and leaves the substance untouched, which is exactly backwards.

The reframe

Make the substance decisions before and during writing, not after. When substance is the first step, it can’t be quietly skipped.

02

Trusting a green audit

Audits score what a script can check in milliseconds: schema validity, heading format, the presence of an FAQ block. None of that checks whether the content holds the attributed data and named sources that actually drive citation.

The reframe

Measure citations, not checkboxes. The audit is hygiene; citation tracking is the scoreboard, and confusing the two is how teams stay proud of a number that doesn’t matter.

03

Publishing what a model can reconstruct without you

If the prose says only what ten other pages say, no amount of formatting earns the citation. A model assembling an answer has no reason to reach for an interchangeable page.

The reframe

Put something on the page a model can’t get anywhere else: original data, a named source, an argument with a position.

04

Orphan statistics and missing sources

A stat with no visible source is a liability, and skipping external citations leaves the single biggest lever unused. If your “73% of buyers” has no link, a model has no reason to cite you over the five pages that attribute the same number.

The reframe

Attribute every number, and cite credible third parties by name. External sourcing is a citation driver, not a footnote.

05

Burying definitions in narrative

When a key term first appears three paragraphs into an anecdote, no machine is extracting that definition. If it only makes sense inside the paragraph it lives in, it’s invisible.

The reframe

Define every key term at first use, in language that still works if the sentence is pulled out of context entirely.

06

Skipping verification

One fabricated stat, one speculative claim presented as fact, one unverified timeline, and the whole piece loses its authority signal as models cross-reference claims across sources.

The reframe

Keep a source log. Trace every claim back to documentation, data, or stakeholder confirmation before it goes live. Accuracy over comprehensiveness.

Honest about the limits

Anyone selling you certainty on AEO is selling you something. The Princeton numbers are from 2024, the engines have changed since, and citation rates move heavily month to month, so treat the magnitudes as directional rather than gospel. What has held up across every content type I’ve applied this to is the order: build the substance, structure it cleanly, then measure whether you’re actually being cited. That part I’m confident about. The exact percentages are softer than the confident playbooks suggest.

Where this came from

Built in a live content function.

If you’re running something similar and it breaks somewhere mine doesn’t, I want to hear about it.