Technical foundation
- Technical audit
- Information architecture
- Technical SEO
- Performance
- Indexing
- Sitemap
- Canonical
- hreflang, where there is more than one language
- Valid structured data matching the visible content
AcquisitionSEO, GEO, AEO and pSEO
Being found matters. Being understood and considered is decisive.
Four fronts on one foundation: SEO to be found in search, GEO to be understood and referenced by artificial intelligence, AEO to hold the place where the answer appears, and pSEO to cover — with data you already own — the long tail nobody can write by hand.
The diagnosis starts with an audit of what exists, not with a keyword list.
Definitions
Four acronyms, one shared foundation
SEO, GEO, AEO and pSEO start from the same foundation — technical structure, specialist content and authority on a subject — and diverge at the destination: a list of results, a generated answer, the block that answers before the click, and the long tail of specific searches.
The practical difference is how the information is consumed. In traditional search, a person weighs a set of options and chooses. In an answer experience, the system delivers a synthesis, and the company appears inside it — or does not. Being comprehensible to whoever synthesises has become as important as being findable by whoever lists.
This does not make SEO obsolete. What feeds answer experiences is, to a large extent, the same indexed content that supports search. Working on GEO, AEO or pSEO with no SEO foundation is building a roof without walls.
The four fronts are not four contracts or four teams. They are different emphases on the same archive: one well-structured page can rank, be cited in a generated answer and hold the direct answer block. What changes is what gets optimised first — and that depends on where the company is actually losing.
- SEO — search engine optimisation
- The set of technical, editorial and authority practices that improve how well relevant pages can be found and interpreted by search engines. It is the foundation of the other three: with no crawling, no indexing and no content that genuinely answers, there is nothing to cite, nothing to summarise and nothing to scale.
- GEO — generative engine optimisation
- Preparing content and digital presence so that the company is understood, considered and referenced in search and answer experiences that use artificial intelligence. It matters because part of all research now ends inside a generated answer: whoever is not understood by the system doing the synthesis does not make it into the synthesis, however well they rank in the list.
- AEO — answer engine optimisation
- Work on format: turning each real question from the audience into an explicit question with a direct, short, verifiable answer right below it, backed by structured data that matches the visible text. It matters because a share of searches is resolved on the screen itself — in an answer block, in "people also ask", or by voice. There you are not competing for a position, you are competing for the answer itself.
- pSEO — programmatic SEO
- Generating pages from a database you own and a single template: one page per real record, each carrying information only it has. It matters because the long tail — specific combinations of need, category, location or application — adds up to a large amount of intent and is impossible to write by hand. It only works when there is real data per page, measured demand and controlled indexing; without all three, it produces empty volume.
The problem
The company exists. Search does not know that.
Search and answer engines do not assess commercial intent. They assess what they can crawl, understand and attribute to someone.
What is usually going on
- A site that does not appear for the searches describing exactly what the company does.
- Content published with no defined search intent.
- Pages on the same site competing with each other for the same subject.
- A technical structure that hampers crawling and indexing.
- A poorly understood brand: the engine does not know what the company is about.
- No topical authority — nothing indicates the company knows the subject.
- Content with no stated authorship and no verifiable sources.
- Pages slow enough to affect both experience and assessment.
- Structured data absent, invalid, or describing something not on the page.
- Audience questions answered only after six paragraphs of preamble.
- The direct answer block for the sector’s queries always held by someone else.
- Dozens of specific searches in the market with no matching page on the site.
- A database of your own — units, plans, applications, cities served — that exists in the system and not on the site.
- Content produced at volume, diluting the site instead of strengthening it.
- An outdated archive, with information that no longer matches what the company does.
- No local SEO, where the buying decision has a geographic component.
- Exclusive dependence on paid media for any visit at all.
What it costs
- Every visitor has to be bought, and the cost rises with competition.
- Smaller competitors take the reference position on the subject.
- The company disappears from generated answers — where part of all searches now ends.
- The question is answered on screen by another source, and the click never arrives.
- Specific demand goes to whoever published the page that answers it.
- Content produced does not accumulate: each piece is born and dies alone.
Having the best answer counts for nothing if nothing on the site lets it be found, understood and attributed to you.
Clearing up the confusion
What GEO, AEO and pSEO are not
These are recent terms, and for that reason they have become wrapping for false promises. Nothing on this list can be delivered.
- A guarantee of citation in generated answers.
- A replacement for SEO.
- A secret schema the search engines never announced.
- A magic file you drop in the site root.
- Manipulation of a language model’s answers.
- Repeating questions across the page to "train" the answer.
- A question block glued to the footer, repeating what the page already said.
- Thousands of pages generated by combining term and city, with no real data behind each one.
- Machine-rewritten text at volume, staged to look like different pages.
- A promise of appearing in every artificial intelligence.
What does exist is foundational work: clarity about what the company does, content that genuinely answers, authorship, sources, well-defined entities, answers in a format the engine can extract, and scale only where there is real data and measured demand to support it.
How we think
Authority is not declared, it accumulates
No technical adjustment turns a site without substance into a reference. What makes an engine — search or answer — treat a company as a source is the accumulation of specific, verifiable, attributed content on a small set of subjects.
That is why the work starts by choosing where it is worth competing. A site that tries to cover everything splits its own authority across dozens of subjects and is a reference in none. Concentrating produces results sooner, and those results fund the expansion later.
Scale comes afterwards, and only with data behind it. The difference between pSEO and a doorway page is not quantity: it is whether each page carries information only it has and a real search looking for it. When that is missing, the right move is not to generate the page — and that recommendation has been given more than once.
The technical side is not incidental: a slow, poorly structured site with invalid structured data prevents good content from being read. But it is a condition, not a cause. Technically fixing a site with no relevant content delivers a fast site that is still invisible.
- One page, one intent
- Two pages competing for the same subject weaken both. Consolidating usually pays off more than producing.
- The answer before the argument
- Someone arriving with a question needs the answer in the first lines. Context comes afterwards, for those who stayed.
- Attribution matters
- Clear authorship, sources and entities are what allow a system to say whose information that is.
- No page without its own data
- If a page exists only because two words were combined, it should not exist. Scale rests on a database, not on a text template.
- Updating is worth as much as publishing
- Content that already holds a position and has lost performance usually pays off more updated than replaced.
Scope
What the area covers
The composition depends on the site’s technical state, the existing archive, the data available and competition on the subject.
Content and intent
- Search intent research
- Subject mapping
- On-page SEO
- Specialist content
- Updating existing content
- Internal links between pages on the same subject
Comprehension and authority (GEO)
- Topical authority
- Stated authorship
- Verifiable sources
- Entity definition
- Semantic clarity
- Consistency between what the site says and what other sources say about the company
Answer format (AEO)
- Collecting the real questions the audience asks
- A direct answer in the first lines of each block
- Question-shaped headings, where that is how the search happens
- Short definitions that can be attributed and quoted
- Tables and lists that can be extracted without the rest of the page
- FAQ and step-by-step structured data repeating the visible text
Scale with your own data (pSEO)
- Assessment of the existing database and what it allows you to publish
- A page model per type of search
- A minimum demand rule: with no measured search, the page is not born
- Unique content per page, coming from the data and not from the template
- Index control and cannibalisation prevention
- Internal linking, pagination and crawl budget for the set
Local reach and reading
- Local SEO, where the decision has a geographic component
- Monitoring of positions and queries
- Visibility analysis
- Tracking of verifiable mentions and references
Not every project uses every item. A technically healthy site with little content needs the opposite of a site with plenty of content and poor structure — and the programmatic front only enters where there is a database to support it.
Process
How the work happens
Fix what blocks, choose where to compete, answer, scale where there is data, and sustain.
Audit
A technical survey, plus archive, indexing, structured data, questions already answered and current position relative to competitors on the subject.
Choosing subjects and questions
Intent research, subject mapping, collection of the real questions the audience asks, and a decision about where it is worth competing — and where it is not.
Fixing the foundation
Structure, performance, indexing, canonical, structured data and resolving pages that compete with each other.
Production, answers and consolidation
New content where it is missing, updates where a position already exists, rewriting blocks so the answer comes before the argument, and internal linking across each subject.
Scale with your own data
Where a database and measured demand exist: page model, generation per record, sample review before publishing and index control. With no database, this step does not happen — and saying so is part of the work.
Monitoring
Tracking impressions, positions, queries, presence in answers and how much of what was published is actually working, with a decision about the next cycle.
What depends on you
- Access to the site and to search and measurement tools.
- Capacity to apply technical fixes, or authorisation for us to apply them.
- Internal sources for specialist content, where the subject is technical.
- Access to your own database, where the programmatic front makes sense.
- Patience with the cycle. SEO has inertia in both directions: slow to rise and slow to fall.
Indicators
What starts being tracked
An isolated position is not a result. The set below is what shows real progress.
- Organic impressions
- How often the site appears in searches across your market.
- Clicks
- How much of that exposure turns into a visit.
- Positions
- Where pages appear, by query and by subject.
- Queries
- Which searches bring the audience — and whether they are the right ones.
- Indexed pages
- How much of the site is actually available to be found.
- Qualified traffic
- Whether people arriving via search match the business profile.
- Organic conversions
- How much of the result comes through this route, without buying media.
- Share of subject
- How much of a subject the company occupies relative to competitors.
- Questions covered
- How many of the audience’s real questions have their own answer on the site.
- Presence in answers
- Whether the page appears in the block that answers before the click, where one exists for the query.
- Long tail covered
- How much of the market’s specific searches have a matching page.
- Yield at scale
- How much of the programmatic set earns impressions — and how much never moves.
- Local visibility
- Presence in searches with geographic intent, where applicable.
- Mentions and references
- Verifiable citations, where they can be identified.
- Topical authority
- Whether the archive as a whole supports the company as a source.
- Content freshness
- How much of the archive stays current and performing.
We do not invent a "GEO score" or an "AEO index". No public methodology supports a single figure for presence in answer engines — inventing one would turn what is still qualitative observation into a number. What is measurable is listed above, and it comes from the official search tools.
When it fits
When these fronts are the right step
They address dependence on media and absence from the research that precedes a decision.
- When the audience researches before buying and the company does not appear in that research.
- When every visitor has to be bought and cost per click only rises.
- When smaller competitors hold the reference position on the subject.
- When your sector’s question is already answered on screen — and the answer is not yours.
- When the company holds a database describing real searches and it has never been published.
- When the site has good content that has never been found.
- When the buying decision has a local component.
- When a new site, a migration or a domain change is being planned.
Limits
When these fronts alone are not enough
This is work of accumulation. It does not solve urgency.
- When the need is a volume of opportunities this month.
- When the product or service is so new that nobody searches for it by name yet.
- When the site cannot be changed technically.
- When there is no way to produce specialist content from internal knowledge.
- When there is neither a database nor measured demand: scaling there produces empty pages, and the recommendation will be not to scale.
- When the page receiving the visitor does not convert — more visits only widen the loss.
When the timeline is short, paid media covers the gap while organic matures. The two areas hold each other up better together than apart.
Questions
Frequently asked questions about SEO, GEO, AEO and pSEO
How long until results appear?
Technical fixes can take effect within weeks. Authority on a subject takes cycles, and depends on the editorial competition around it and on what already exists on your site. The diagnosis estimates that from the real state, not from a standard timeline.
Do you guarantee first position?
No, and nobody can. Position depends on factors no supplier controls: competition, algorithm changes and the engine’s own assessment. What is controllable is the work — structure, content, authority and clarity.
Does GEO actually work, or is it just a new name?
The name is new; the work is largely the same well-executed SEO, plus attention to semantic clarity, authorship attribution and entity definition. What does not exist is a technique guaranteeing citation in a generated answer — anyone promising that is selling what they cannot deliver.
What is AEO, and why does it matter if nobody clicks?
AEO is answer engine optimisation: preparing each real question from your audience to be answered directly, briefly and verifiably, with structured data repeating the visible text. It matters precisely because a share of searches is resolved without a click — in an answer block, in "people also ask", or by voice. Appearing there builds recognition and tends to pull the next query, which is the one closer to a decision. When the click does come, it arrives better qualified, because the basic question has already been answered.
Is pSEO not exactly what you say does not work?
What does not work is generating pages by combining words, with no information of their own — that still holds. pSEO done with judgement is the opposite: there is a real database behind it, each page carries a fact only it has, search demand is measured before publishing, and indexing is controlled. When those four requirements are missing, the recommendation is not to generate the pages, and it has been given more than once.
Would it not be better to publish many more pages?
Almost never, and never on its own. Volume without depth dilutes authority and creates pages that compete with each other. Scale only holds up when each page comes from a real record in a database you own and answers a search that exists — beyond that, the site grows in addresses and shrinks in authority.
Do I need a blog to do SEO?
Not necessarily. Many companies gain more by organising and deepening the pages that already describe what they sell than by opening a blog nobody will have the stamina to maintain.
Is local SEO included?
Where the buying decision has a geographic component, yes — it is one of the areas in scope. Where it does not, including it would be work with no return.
Before choosing keywords, it helps to know what stops the site being read.
The diagnosis starts from an audit of what exists — structure, archive, indexing, the shape of the answers and the data available — and points out which of the four fronts pays off first.
No proposal is sent before the diagnosis conversation.