AEO vs SEO: Same Crawl, Different Scoreboard

No — Answer Engine Optimization is not SEO with a new name. You still need crawlable pages, clean entities, and trustworthy sources. What changes is the scoreboard: instead of optimizing mainly for blue-link rank and click-through rate, you optimize for inclusion in synthesized answers — citations, accurate brand facts, and share of voice across prompt panels.

Google’s own line is the grain of truth behind “it’s still SEO.” Generative AI features on Google Search are rooted in core Search ranking and quality systems. Eligibility is still “indexed and snippet-eligible” — no extra technical requirements. Polarized takes online either declare SEO dead or treat that sentence as permission to keep a 2019 dashboard. Both miss the operator view.

This spoke maps what actually changes — unit of win, KPIs, briefs, budget, and failure shape — under the Answer Engine Optimization playbook. It does not rebuild the five-layer method or the 90-day sequence. Those live in the pillar. The definition of the new scoreboard lives in what AI visibility is.

The short answer

• SEO wins documents a place on a results list. AEO wins passages a place inside an answer.
• Crawl, indexation, performance, and entity hygiene remain shared infrastructure. Google says so in public.
• New AEO-specific work: prompt panels, citeable structure, corroboration, hallucination cleanup, multi-engine measurement.
• AEO does not replace SEO. Weak SEO usually means a thin candidate pool for answer engines.
• Tip budget toward AEO when buyers already ask recommendation questions in ChatGPT, Perplexity, or Google AI Overviews — not when you have never measured those surfaces.

What “same crawl, different scoreboard” means

Search systems and answer systems both need to fetch your URLs. That is the shared crawl layer: HTTP 200s, sane robots rules, indexable canonicals, readable HTML, stable entity names. Google’s AI-features doc is blunt: to show as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible for a Search snippet. There is no second door.

The scoreboard diverges after retrieval.

Layer	SEO primary KPI	AEO primary KPI
Visibility	Rank / impressions	Citation rate / mention rate
Traffic quality	CTR / sessions	Qualified AI referrals + brand-search lift
Trust	Backlinks / E-E-A-T signals	Corroboration across domains + fact consistency
Content job	Cover the query intent	Survive compression into a 60-word answer
Failure mode	Drop from page one	Cited wrong, skipped, or competitor recommended

If your weekly dashboard only shows average position, you are flying half-blind on AI surfaces. AI visibility is the inclusion scoreboard — named, cited, or both — not a renamed rank report.

Pew’s July 2025 short read on March 2025 browsing data is the click-side receipt: when an AI summary appeared, users clicked a traditional result in 8% of visits versus 15% without one, and they clicked a link inside the summary in 1% of visits (Pew Research Center). Rank can hold. Clicks still fall. That is a scoreboard change, not a crawl change.

What’s the difference between AEO and SEO?

SEO asks: which document should I show for this query? AEO asks: which facts and brands should appear in the answer I write?

Semrush draws the same split in operator language: SEO aims to rank in traditional results (and Google’s AI-powered SERP features) to drive traffic; AEO aims to get cited or mentioned inside generated answers (Semrush: AEO vs SEO). Shared toolkit. Different acceptance test.

Practical differences operators feel week to week:

#	Dimension	SEO default	AEO default
1	Unit of win	Page position	Passage inclusion
2	Query surface	Keyword SERP	Conversational prompt, often multi-intent
3	Attribution	Click	Citation chip / named recommendation
4	Corroboration weight	Links help both	Off-site agreement matters more when engines synthesize
5	Failure shape	Traffic dip	Wrong founding year, wrong pricing, competitor named instead of you

Same pieces show up in both columns — schema, content, PR. The brief and the acceptance criteria change.

Worked example — one money page, two scoreboards:

Page	SEO outcome	AEO outcome	What you do next
/services for a studio	Rank 3, steady clicks	ChatGPT names a competitor and invents your starting price	Fix the quote-ready lead + schema/About price string; re-run the panel
Local HVAC service area	Local pack + #2 organic	Overview cites a directory, not you	Corroborate NAP on the directory; do not only tweak the title tag
B2B comparison URL	Rank 8, low CTR	Cited as the source chip, almost no click	Count the citation as a win; build a branded next step for the people who do click

ChatGPT Search makes the attribution shift concrete. When search runs, answers may include inline citations plus a Sources panel. There is no “position 3.” There is “named,” “cited,” “absent,” or “wrong.” Log those four outcomes. Do not invent a ChatGPT rank.

Does AEO replace SEO?

No. Treat AEO as a layer on top of a working search foundation, not a replacement religion.

Google’s May 2026 optimization guide opens with the sentence operators screenshot: SEO best practices continue to be relevant because AI Overviews and AI Mode retrieve from the Search index with RAG, then synthesize. If you are not in that index with a snippet, you are not in the candidate pool. That is not a slogan. It is the eligibility rule.

• Indexation and technical SEO still pass basic hygiene
• Priority money pages still compete for classical intent where clicks matter
• AEO budget funds answer-first pages, entities, corroboration, and multi-engine logging
• You stop pretending “we rank #3” equals “ChatGPT recommends us”

Teams that “pivot fully to AEO” and abandon crawl health usually discover their pages never enter the candidate pool. Teams that refuse AEO KPIs keep celebrating ranks while buyers get competitor names in chat.

I have been SEO-certified since 2021. The crawl work did not get optional. The weekly question did.

Replacement test: if you deleted every AEO ticket and kept only technical SEO, would ChatGPT Search start citing you? No. If you deleted every SEO ticket and kept only prompt-panel content, would Google still retrieve you? Often no. You need both. You do not need two religions.

What does Google mean by “still SEO”?

Google means: AI Overviews and AI Mode are Search features, not a second internet with a weekly rulebook. They use query fan-out and RAG against the existing index. They trigger only when the system judges the overview additive to classic Search — and often do not trigger at all.

What it does not mean:

Misread	What Google actually published	Operator correction
“Keep the old rank report and you are done”	A dedicated Generative AI performance report now tracks impressions in AI Overviews and AI Mode	Add that report. It is impressions, not a citation-rate substitute
“Meta-description A/B tests are the AI program”	People-first, helpful, reliable content is still the quality bar	Change the page shape so a 60-word answer can lift a true sentence
“AEO is a Google-invented tactic pack”	Google tells site owners they can ignore many AEO/GEO “hacks” for Google Search	Believe them for Google. Do not pretend ChatGPT Search is Google Search
“Special AI files unlock Overviews”	Google Search does not use llms.txt as a ranking or inclusion file	Do not sell an llms.txt as an Overview ticket

Keep the foundation. Change the scoreboard and the content shape. If leadership quotes “Google said it’s still SEO” to kill measurement, they quoted a crawl sentence and skipped the reporting sentence.

How does query fan-out change the content job?

Classic SEO briefs one primary keyword and a cluster of secondaries. Fan-out breaks one prompt into concurrent sub-queries, then retrieves pages for those sub-queries, then writes one answer. Google’s own example: “how to fix a lawn that’s full of weeds” can fan into herbicides, non-chemical removal, and prevention (Google AI optimization guide).

That changes the job without changing the crawl.

Old SEO habit	What fan-out does to it	What to do instead
One URL, one head term	The model may retrieve you for a sub-query you never “targeted”	Map the real sub-questions buyers ask; answer each in a stand-alone section
Exact-match title obsession	Google says you do not need a page per synonym variation	Do not spawn thin variants to chase fan-out strings — that is scaled-content risk
Word-count floors	Retrieval is passage-level; length is not the win	Lead with a quote-ready answer; put the proof in a table or steps
“Cover the topic” fluff	Commodity roundups are easy to compress into a generic sentence	Add the first-hand fact only you can defend

Google also says you should not manufacture a page for every fan-out variation just to manipulate responses. That is scaled content abuse, not AEO.

Checklist for a money page that has to survive fan-out:

• The first 2–4 sentences answer the spoken question out loud
• Each H2 can be lifted without the rest of the page
• One table, numbered procedure, or decision list per major section
• Named entities spelled the same way as schema, About, and directories
• No second URL that repeats the same answer with a synonym title

The crawl did not change. The unit the model lifts did.

Why isn’t ranking a citation?

Because retrieval and synthesis are not the same step as ranking a blue link.

Ahrefs’ July 2025 study of 1.9 million AI Overview citations found 76% of cited pages also sat in the top 10. Their March 2026 update — 863,000 keyword SERPs, 4 million AI Overview URLs — found 38% of cited URLs also appeared in the first 10 blocks (Ahrefs: 38% of AI Overview citations). Overlap shrank. Rank is still useful. It is a weaker proxy than it was a year earlier.

An earlier Ahrefs pass on one million Overview-triggering keywords found a Spearman correlation of 0.347 between top-10 ranking and a top-3 Overview citation — and even a #1 page appeared in those top-three cites only about half the time (Ahrefs: does ranking higher mean you’ll get cited?).

Evidence	What it proves	What it does not prove
Google: indexed + snippet-eligible	You need SEO hygiene to be eligible	Eligibility is not inclusion
Ahrefs 76% → 38% top-10 overlap	Rank and citation used to travel together more often	“Rank #1” is not an Overview ticket
Ahrefs ~50% of #1s in top-3 cites	Strong SEO helps; it does not close the loop	A rank report can replace a prompt panel
Pew 8% vs 15% clicks	Overviews change the click economy	Your brand was named in the answer

If you only track position, you will miss the week you stayed #2 and disappeared from the Overview, and the week you sat #14 and got cited on a fan-out passage. Those are different failures. They need different tickets.

How do KPIs actually change?

You do not average rank and citation rate into one vanity score. You run two columns on one page.

Search Console now splits the Google side. The Generative AI performance report (announced June 2026) shows impressions for AI Overviews and AI Mode by page, country, device, and date (Search Central blog). That is Google’s official AI-surface column. It is not ChatGPT. It is not accuracy. It is not share of voice against a competitor set you chose.

KPI	SEO column	AEO column	Where it lives
Primary	Top-10 share on money queries	Citation rate on a frozen prompt panel	Rank tracker + human-reviewed log
Secondary	Organic sessions / leads	Mention rate + SOV vs named competitors	Analytics + panel sheet
Google AI surface	Web Performance (clicks, CTR)	Gen AI impressions (Overviews / AI Mode)	Search Console
Trust / risk	Indexation and ranking losses	Wrong facts, competitor SOV, hallucinated prices	Accuracy review
Non-goal	Vanity Domain Authority chase	“Rank #1 in ChatGPT” promises	Slideware

SparkToro’s 2026 Similarweb-panel read puts the click economy in a longer arc: 68% of U.S. Google searches in the first four months of 2026 ended without a click to anywhere (SparkToro). Panels differ across years; do not treat 68% as a Google-official number. Treat it as the reason CTR-only success stories age badly.

• Classic Performance report still reviewed weekly
• Generative AI impressions reviewed on the same cadence
• Prompt panel logged for ChatGPT Search and any other engine sales actually hears
• Accuracy incidents opened when the answer is wrong, even if you were named

One-week logging procedure that keeps the columns honest:

1. Freeze the prompt list. Do not add “who is [you]?” vanity prompts to the win rate.
2. Run each prompt once in ChatGPT Search and once on the Google queries that actually trigger an Overview.
3. Score four fields: mentioned (y/n), cited (y/n + URL), fact-accurate (y/n), competitor named (list).
4. Pull GSC Web Performance and, if present, generative AI impressions for the same URLs.
5. Put both columns on the one-pager before you write new content.

If finance only funds the left column, say you are not funded for AEO. If they only fund the right column and ignore crawl health, push back equally hard.

What work is shared vs unique?

Use this split when you defend budget or write a roadmap. Shared work is funded once. Unique work gets its own tickets.

Shared (keep / fund once)	SEO-heavy	AEO-unique
Crawl / index / canonicals	Keyword research for SERPs	Prompt panel design + weekly runs
Core Web Vitals (good enough)	Classic link building	Citation-gap maps across engines
NAP / entity consistency	Title / meta CTR tests	Hallucination / wrong-fact cleanup
Honest schema that matches visible text	Thin-page consolidation	Brand briefing files other engines may read — not a Google Overview ticket
Topical clusters	Local pack / Business Profile hygiene	Answer-first passage engineering

Google is explicit that structured data is not required for generative AI search and that there is no special schema.org type that unlocks Overviews. Keep schema for the SEO reasons you already had — rich-result eligibility where it still exists, and unambiguous facts. FAQPage remains a valid type. Google’s FAQ rich result is not the reason to write FAQ H3s. Visible Q&A that matches buyer questions is.

Ticket rewrite — same URL, different definition of done:

Old SEO ticket	AEO-aware ticket
“Add the keyword to H1 and meta”	“Answer the spoken question in sentence one; keep the H1 honest”
“Build 5 links this month”	“Get one independent page to repeat the same founding year / price / category”
“Publish 8 supporting posts”	“Kill two commodity posts; deepen one comparison table”
“Improve CTR 0.4 points”	“Re-test prompt IDs 12–18; close the accuracy incident if the price is still wrong”

Semrush (or equivalent) still earns its seat for SERP and competitive discovery. It does not replace a human-reviewed AI answer log. Vendor AI-visibility scores are amplifiers. They are not the ritual.

How does this change content briefs?

Old brief (SEO-only): primary keyword, secondary terms, word count, competitor URLs, title ideas.

That brief can still ship a page that ranks and still lose the answer. The AEO-aware brief adds fields that change the acceptance test.

Field	Why it exists	Fail if missing
Primary question (spoken)	Models quote answers to questions, not keyword clouds	Title is a noun pile
Quote-ready definition (≤60 words)	Survives summarization	Lead buries the take in paragraph four
Structured element required	Tables and steps get extracted	Wall of prose, no lift point
Entities to name consistently	Stops synonym drift	“WS,” “Spurlock,” and “the studio” all mean you
Corroboration ask	Which external page should agree with this claim?	Unique statistic with no second source
Prompt panel IDs to retest	Proof the brief moved a KPI	“We published” is the only success metric

Google’s quality bar still applies: helpful, reliable, people-first content, including E-E-A-T signals. The AEO add is not “write for robots.” It is “write a sentence a robot can lift without inventing your price.”

If the brief cannot name how you will re-measure inclusion, it is still a rank brief wearing an AEO hat.

1. Write the spoken question at the top of the ticket
2. Draft the 60-word lead before the outline
3. Require one table or procedure in the outline, not in the “nice to have” comment
4. List the three entity strings that must match the rest of the site
5. Attach the prompt IDs you will re-run in two weeks

What should an SEO retain — and who owns AEO?

Keep the skills. Change the acceptance tests. Usually you do not hire a second team.

Skills that transfer:

• Technical crawl diagnostics
• Intent mapping (extend it into prompt types)
• Internal linking and cluster thinking
• Digital PR — retarget for citeable corroboration, not only Domain Rating
• Local / Business Profile hygiene where it applies

Add these to existing tickets:

• Lead answer in the first 2–4 sentences
• At least one table, checklist, or numbered procedure per major section
• FAQ H3s that match real buyer questions
• Fact-sheet alignment with schema and About

Company shape	Recommended ownership	Red flag
Solo / small studio	SEO or content lead runs the panel; founder reviews accuracy monthly	Renaming the freelancer “Head of AEO” with no log
Mid-market	SEO lead owns AEO KPIs; content writes answer-first; PR owns corroboration	A separate agency that cannot show crawl competence
Enterprise	Visibility pod (SEO + content + brand) with a shared prompt panel; legal on high-risk claims	A title without a weekly cadence

A separate “AEO agency” that cannot show SEO competence is a red flag. A separate headcount title without a measurement cadence is theater. Google Search still rewards people-first pages that meet technical requirements. ChatGPT Search still rewrites prompts and cites sources. Someone has to own both columns.

When should budget tip toward AEO tactics?

Tip when two or more of these are true. Stay SEO-weighted when you have no crawl foundation, no measurement ritual, and no evidence buyers use answer engines for your category. Guessing is not a tip signal.

Signal	Why it counts	What you fund next
Sales hears “I asked ChatGPT / Perplexity who to use”	Demand already moved	Prompt panel + competitor SOV + five answer-first money pages
Informational CTR falls while impressions hold	Overview or chat displacement	GSC gen-AI impressions + Pew-style click humility + page-shape rewrites
Competitors appear in AI answers and you do not	Citation gap, not a rank gap	Citeable comparisons + corroboration, not another title-tag test
Brand facts in AI answers conflict with your site	Accuracy incident	Truth layer: site, schema, directories, then re-test
Competent SEO, wrong next dollar	Diminishing rank returns	Measurement + truth layer before more content volume

Google’s own product PDF says AI Overviews reach more than 1.5 billion users. That is a surface size argument, not a “your category triggers Overviews” argument. Check your queries. Check your sales calls. Then move money.

Stay-weighted checklist — if three or more are true, do not tip yet:

• Indexation or robots issues still open
• No money-query rank baseline you trust
• No one on the team can run a 25-prompt panel this week
• Sales has never heard an AI-referred origin story
• The site still contradicts itself on price, location, or offering

A tip without those five boxes is a rebrand. A tip after them is a scoreboard change.

Where does GEO fit without three religions?

GEO — Generative Engine Optimization — is the label teams use when the engine synthesizes prose (Overviews, ChatGPT with search, similar products). At Spurlock Studios, GEO sits inside AEO; AEO sits beside SEO. This post owns the rebrand objection, not the glossary war.

Google’s optimization guide mentions AEO and GEO as common internet labels and then tells site owners that many associated “hacks” are not how Google Search works. That is a Google-Search statement. It is not a claim that ChatGPT Search, or any other answer product, scores pages the way a 2019 rank tracker does.

Rule of thumb: one strategy doc, three KPI columns (SEO / AEO / GEO-surface notes), one roadmap.

Label	Use it for	Do not use it for
SEO	Eligibility, classical SERPs, GSC Web Performance	Claiming ChatGPT “rank”
AEO	Inclusion in answers across engines you actually care about	A second agency with no crawl skills
GEO	Notes on synthesis-heavy surfaces inside the AEO column	A third religion with a third retainer

If a vendor needs three acronyms to sell one dashboard, buy the ritual first.

Failure mode: the rebrand-only program

What breaks: leadership renames the SEO team to “AEO,” buys a dashboard logo, and ships the same thin listicles.

What it costs: three months of budget, no citation lift, and a board that now thinks “AI search is fake.”

What you do instead: keep SEO hygiene, ship one prompt panel, fix the top five fact conflicts, rewrite five money questions answer-first, then expand. Sequence beats slogans.

Symptom	What is actually broken	First fix
New title, old tickets	Acceptance tests never changed	Add lead-answer + structure + retest IDs to the brief
Dashboard, no panel	Vendor % with no frozen prompts	Freeze 25–40 prompts and 3 competitors this week
“We published 40 AEO blogs”	Volume without a unique fact	Kill the commodity roundups; keep the five pages with receipts
Rank held, sales hears a competitor	Scoreboard was only SEO	Log the exact prompt and the exact wrong answer
llms.txt as the whole program	Google Search does not use that file	Treat it as optional briefing, not an Overview lever

Hundreds of production sites taught me the cheap version of this failure: a rename is free. A weekly log is not. Fund the log.

What should the one-pager look like?

One page. Two columns. No thesaurus. If leadership only wants the left column, you are not funded for AEO yet — say so.

Row	SEO column	AEO column
Primary KPI	Top-10 share on money queries	Citation rate on the panel
Secondary	Organic sessions / leads	Accuracy incidents open
Google-only view	Web Performance clicks / CTR	Gen AI impressions
Risk	Indexation / ranking losses	Wrong facts / competitor SOV
Ship this month	Rank-protecting fixes	Truth layer + 5 answer pages
Non-goal	Vanity DA chase	“Rank #1 in ChatGPT” promises

Budget split for teams that already have competent SEO. Not a law. Revisit after each re-baseline. Frozen percentages become superstition.

Situation	Suggested next-dollar split
No AI panel ever run	70% measurement + truth layer / 30% SEO maintain
Panel shows accuracy fires	60% fact / corroboration / 40% content
Panel shows competitor citations, you absent	50% citeable content / 30% PR / 20% tech
Strong citations, weak classical leads	Shift back toward SEO conversion paths

Pew’s click drop and SparkToro’s zero-click trend are reasons to measure the right column. They are not reasons to abandon the left. People still click. They click less often when an overview is sitting on the query.

What changes in the first 30 days?

Not a new org chart. Not the pillar’s 90-day build. A transition stack so the two scoreboards exist in the same week.

1. Freeze a 25–40 prompt panel and three competitors
2. Baseline ChatGPT Search (citations / Sources), plus priority AI Overviews
3. Turn on and read the GSC generative AI report if the property has it
4. Reconcile brand facts (site, schema, directories)
5. Rewrite five high-intent pages answer-first with structure
6. Open a citation-gap backlog for PR / partners
7. Report SEO KPIs and AEO KPIs on the same one-pager

That is the practical stack view. Depth on entities, schema, clusters, and the longer loop lives in the playbook. This spoke is done when the team can explain — in one sentence — what changed versus last year’s SEO program.

The sentence is: we still crawl and compete for clicks; we now also log whether the answer includes us, cites us, and gets the facts right.

Done looks like this, not like a rename:

• Two columns on one page, reviewed the same week
• Five money pages with a spoken question and a lift-ready lead
• Accuracy incidents have owners, not screenshots in Slack
• Next dollar assigned from the panel, not from the acronym debate

FAQ

What does Google mean by “still SEO”?

Google frames AI Overviews and AI Mode as Search features built on crawl, ranking, and helpful-content systems — not a brand-new internet. Eligibility is still indexed, snippet-eligible pages. Operators still need extractable answers, consistent entities, and a second measurement column beyond classical rank reports.

How do KPIs differ (rank/CTR vs citation/SOV)?

SEO tracks positions, clicks, and conversions from SERPs. AEO tracks citation rate, mention rate, share of voice on a prompt panel, and accuracy of brand facts inside answers. Google’s generative AI report adds Overview and AI Mode impressions; it does not replace a human-reviewed answer log. Run both columns — do not average them into one vanity score.

Where does GEO fit?

GEO names optimization for generative engines that synthesize answers. Treat it as a subset of AEO in planning docs so you do not fund three agencies for one roadmap. Google’s public guidance is about Google Search; other answer products still need their own inclusion logs.

Do I need a separate AEO team?

Rarely. Name an owner, attach AEO KPIs to the SEO or content lead, and add PR for corroboration. Split teams only when volume and risk justify a visibility pod with a shared panel. A vendor that cannot diagnose crawl issues is not an AEO team.

What should an SEO retain if we add AEO?

Keep technical SEO, intent mapping, clustering, and PR. Change briefs and acceptance tests so pages win passages, not only positions. Retain Semrush-class SERP tooling. Add a human-reviewed AI answer log and the Search Console generative AI impressions view.

How does this change content briefs?

Briefs must include a spoken primary question, a quote-ready lead answer, required structured elements, entity spelling, a corroboration ask, and the prompt IDs you will retest. Word count alone is not a strategy. If you cannot name the re-measure, you wrote a rank brief.

CTA

Stop arguing the acronym. Measure both scoreboards, then fund the gaps.

Lane overview: /visibility. Book a visibility audit when you need an external prioritization pass on SEO vs AEO spend.