Google ranks you. ChatGPT decides whether to mention you at all.
SEO, AEO, and GEO are three different games played on the same website. Most firms are optimized for one of them and invisible in the other two — and have no way to tell which.
Three acronyms, three definitions of winning
The confusion in this market comes from treating these as synonyms. They are not. Each one has a different unit of success, and optimizing for one does not automatically deliver the others.
SEO
Search Engine Optimization
Unit: the page
- What you win
- A position in a ranked list of links.
- What it runs on
- Crawlability, intent match, internal linking, earned links, page experience, technical health.
- What failure looks like
- You're on page two. Someone clicks a competitor instead.
AEO
Answer Engine Optimization
Unit: the passage
- What you win
- The extraction — your words lifted into the answer box.
- What it runs on
- Question-shaped headings, the answer in the first 40–60 words, tight question-to-answer proximity, FAQ and HowTo structured data.
- What failure looks like
- You rank third, but the firm ranked seventh owns the snippet above you.
GEO
Generative Engine Optimization
Unit: the fact
- What you win
- Inclusion in a synthesized answer — with or without a link.
- What it runs on
- Entity clarity, off-site corroboration, access for AI crawlers, and claims that survive being paraphrased.
- What failure looks like
- A prospect asks for three firms like yours. The model names three. You are not one of them, and no report will ever tell you why.
Your traffic is down and your rankings didn't move.
That is the signature of an answer surface eating the click before it reaches you. It is also the one problem a conventional rank-tracking report is structurally incapable of showing you — because nothing it measures has changed.
Six principles that hold across all three
Tactics churn. These don't. Every finding in a SEO Doctor audit traces back to one of them.
-
Retrievability precedes everything
Different engines send different crawlers with different rules. Googlebot is not OAI-SearchBot is not ClaudeBot is not PerplexityBot. A robots.txt line or a server-level block isn't a ranking penalty — it's a total loss, and it is invisible from inside your CMS.
-
Be unambiguous about who you are
One canonical entity. Consistent name, address, and phone everywhere. A clean schema graph with real, resolvable identifiers, and profile links that actually exist. Ambiguity was survivable in classic SEO. In generative search it's disqualifying — a model that can't tell which firm you are will pick one it can.
-
Optimize the chunk, not the document
Retrieval systems operate on passages, not pages. A well-argued page that buries its key claim in paragraph nine won't be retrieved. The answer goes first; the supporting argument goes after it.
-
You cannot self-declare authority
Classic SEO rewarded on-page assertion plus links you could build. Generative engines weight independent corroboration — bar directories, review platforms, professional associations, press, and people citing you without being asked. Off-site work matters more now, not less.
-
Specificity beats fluency
Named numbers, dates, jurisdictions, credentials, and quotable statements get lifted into answers. "Trusted partner delivering tailored solutions" gets skipped by every engine and every human. Fluent copy that says nothing specific is the single most common finding we write up.
-
Measurement changes shape
Rank becomes presence becomes citation share. There is no position one inside a generated paragraph. The honest question is: are you mentioned, are you described accurately, and does it happen for the queries that actually precede a phone call?
What a diagnostic actually looks at
Every audit covers all three layers, because you cannot tell which one is failing until you've checked them separately.
- Crawler access by engine, at both robots.txt and server level
- Log evidence of which AI crawlers have actually reached you
- Entity consistency across your site, directories, and profiles
- Schema graph structure and identifier integrity
- Passage-level answer readiness on money pages
- Question coverage against what prospects actually ask
- Live prompting across the major answer engines
- Accuracy of what engines currently say about you
- Off-site corroboration and citation footprint
- Conventional technical SEO health and indexation
- Intent match between your pages and real queries
- Conversion path from answer surface to contact
Every finding is labeled with how much we actually know
Large parts of SEO are settled and testable. Much of AEO is directional. A meaningful share of what circulates as GEO best practice is extrapolation from a single 2024 academic paper plus a great deal of vendor folklore. We refuse to flatten that distinction, so we label it instead.
Verified
We observed it directly and can show you the evidence.
Directional
Strong supporting evidence and consistent industry results, but not individually proven on your site.
Pending
Requires access, data, or a time window we don't have yet. Named, not hidden.
Unconfirmed
Plausible and worth knowing, but we cannot stand behind it. Priced accordingly in your priority order.
We diagnose. We don't take the keyboard.
SEO Doctor does not sell retainers, implementation, or lead generation. You get a flat-fee diagnostic that names every problem in priority order, a remediation work order your own webmaster or developer can execute, and re-audits at 30, 60, and 90 days to confirm the fixes actually landed.
You also get 90 days of unlimited 15-minute coaching calls, because the person doing the work will have questions and you shouldn't have to buy a retainer to get them answered.
The reason is structural: a firm that both grades the work and does the work has an incentive problem. We removed it.
Questions we get asked
Is GEO just SEO with a new name?
No. They overlap at the foundation — a site an engine can't reach can't win either game — but they diverge sharply above that. SEO competes for a link position; GEO competes for a mention inside a generated paragraph where no link position exists. Entity clarity and third-party corroboration matter far more for GEO than for classic ranking, and the measurement is entirely different.
Do I need all three, or can I pick one?
You need to know which one is failing before you spend anything. In practice most firms we audit are reasonably competent at conventional SEO, thin on answer readiness, and completely unmeasured on generative visibility. That mix is common enough that we check all three by default rather than asking you to guess.
Can anyone guarantee that ChatGPT will recommend my firm?
No, and you should treat any such guarantee as disqualifying. Generative answers are non-deterministic and change between sessions, models, and model versions. What can be done is measurable: remove the access blocks, resolve the entity ambiguity, build the corroboration, then re-test on a schedule and track whether mentions increase.
How is this different from a free SEO audit tool?
Automated tools grade what they can count. They do not prompt live answer engines, evaluate whether an engine describes your firm accurately, check server-level crawler blocks that never appear in robots.txt, or distinguish a finding they verified from one they inferred. The difference isn't volume of findings — it's whether the findings are true.
Who do you work with?
Boutique law firms in estate, family, and business practice; fee-only financial advisors; CPA and accounting firms; and established real estate teams. We are a poor fit for solo agents and for anyone whose goal is purchased lead volume rather than a website that earns its own visibility.
Find out which of the three is actually costing you.
Flat fee — $1,500 / $2,500 / $3,500 by scope · No retainer · No implementation