Skip to content
busyless
Book a call
Answer engine optimization

AEO agency work. Without the agency.

Getting quoted by ChatGPT, Google's AI Overviews, Perplexity and Gemini is not a channel you buy by the article. It is a system: what the machines can tell about you, what they can lift from you, who else vouches for you, and whether any of it moved.

I own that system as one operator, and I built the tracker that reads it.

Book a 30-min callSee the actual work
answer engine – live response
Composing
ChatGPTAI OverviewsPerplexityGemini

Buyer prompt

“who should we use for content leadership if we're a 20-person B2B SaaS?”

For a team that size, most people choose between three routes1 – a boutique shop, a fractional lead, or a first in-house hire2.

The names that come up repeatedly arethree vendors34.

Sources used4 of 5
1reddit.comCommunity thread
2g2.comReview category
3competitor.com/blogComparison page
4youtube.comWalkthrough
–yourbrand.comNot retrieved

Illustrative. Your brand is in the market this answer describes, and absent from the sources it was built from.

Every surface a buyer might ask, scanned as one set

  • ChatGPT

    Search + browse

  • Google AI Overviews

    In the SERP

  • Google AI Mode

    Full answer surface

  • Perplexity

    Own crawler

  • Gemini

    Grounded answers

  • Copilot

    Bing-backed

The shift

The page you optimised for is not the page they see.

For twenty years the deal was simple. You earned one of ten positions, and the reader chose from the list. That deal is being renegotiated in front of us.

Now a machine reads a handful of sources, writes the answer itself, and names the few it used. You are no longer competing for a position. You are competing to be one of the sources, and the losing move is invisible: you do not drop to page two, you simply are not in the paragraph.

The page you optimised forTen slots
1
2
3
4
5
6
7

You competed for a position. The reader picked from a list.

The page they getOne answer
redditg2competitoryoutube
1
2

You compete to be a source. The machine picks, then summarises.

The mechanism

AEO is not a new channel. It is a new reader.

Nothing here asks you to publish somewhere new. It asks you to write for something that reads differently from a person and decides differently from a search engine.

Six things decide whether you get quoted. Most answer engine optimization services sell you the first two, because they are the two that fit neatly into a monthly deliverable.

  • Can it tell what you are?

    EntitiesSchemaConsistency

    Before a model can recommend you it has to resolve you: this company, this category, these products, this founder, and not the similarly named business two states over. Most sites never state it in a form a machine can read.

  • Is there a sentence worth lifting?

    PassagesStructure

    Answers are assembled from passages, not pages. A claim buried in paragraph nine of a 2,000-word post is not retrievable. The same claim, stated once and cleanly under a heading that matches the question, is.

  • Does anyone else say it?

    RedditReviewsVideoPress

    Models weight corroboration heavily, and the sources they lean on are rarely your own site. Community threads, review categories, video and editorial coverage carry disproportionate weight, which is uncomfortable and true.

  • Can the crawler actually reach it?

    robots.txtRenderingFeeds

    Several of these engines run their own crawlers, and plenty of sites block them by accident or ship their body text through JavaScript that never gets executed. This is the cheapest defect on the list and the most common.

  • Is it current and specific?

    FreshnessFiguresDates

    Vague, undated and unattributed copy loses to a page that commits to a number and says when it was true. Specificity is what makes a passage safe for a model to quote, because it is checkable.

  • Do the engines agree with each other?

    ConsensusCoverage

    They frequently do not. The same question put to two engines can return two different sets of sources, because they index differently and retrieve differently. Optimising for one and declaring victory is the standard mistake.

Read this before you sign anything

What an AEO agency cannot promise you.

This category is eighteen months old and selling hard. Almost every AEO agency page you will read today promises visibility in AI answers as though it were a rankings report with a new wrapper. Four things make that untrue, and any provider worth paying will tell you all four before you pay them.

  1. 01

    There is no rank, so there is no ranking report.

    An answer is generated per query, and often per user. Two people asking the same thing can get different sources. Anything sold to you as position tracking for AI is a measurement convention somebody invented, not a position that exists.

  2. 02

    Nobody controls the model.

    This work is influence, not instruction. We change what the retrievable record about you says and how easily it is found. What a model does with that on any given day is outside anyone's hands, including the vendors selling certainty about it.

  3. 03

    The referral data is thin, and sometimes absent.

    A lot of these answers end the journey rather than passing a click. Attribution here is genuinely partial, and the honest read pairs citation share with branded search and assisted conversions rather than pretending a clean line exists.

  4. 04

    Half the work is not on your website.

    If the sources being cited are a community thread and a review category, then on-site changes alone will not move it. Anyone quoting you a pure on-page scope has scoped the comfortable half of the problem.

None of that makes the work not worth doing. It makes the measurement honest, which is the only way you will ever know whether it was.

The work

Six workstreams, and none of them is a content calendar.

This is what an AEO audit turns into once somebody has to actually do it. Four of the six start in the first three weeks because they unblock everything else.

  1. 01

    Weeks 1–2

    Establish the entity

    Get the machine-readable facts about your business consistent and complete: organisation and person markup, the profiles that corroborate them, product and pricing statements that do not contradict each other across the site. This is the entity optimization layer, and nothing downstream works without it.

    Entity graph2 unresolved
    Organizationresolved
    sameAs · 6 profilesresolved
    Person · foundermissing
    Service · 3 offersconflicting
  2. 02

    Weeks 2–6

    Make the claims liftable

    Rework the pages that already carry authority so each answers one question in one place, in a sentence a model can quote without editing it. This is a rewrite of structure, not a rewrite of substance, and it usually improves the page for humans too.

    Passage rewritebefore → after

    Buried

    Lifted

    The claim moves to the top of its own section, in one liftable sentence.

  3. 03

    Ongoing

    Build the corroboration

    Go and be present where the citations actually come from. Community threads, review categories, video, podcasts and earned editorial coverage. We do not buy placements, because a bought mention is exactly the kind of thin signal these systems are being tuned to discount.

    Off-domain surfaces2 live
    Community threadsActive
    Review categoriesActive
    Video walkthroughsQueued
    Editorial mentionsQueued

    Nothing here is a placement we bought. Earned, or it does not count.

  4. 04

    Weeks 1–3

    Clear the retrieval path

    Audit what the AI crawlers can reach and read: robots directives per user-agent, server-rendered body text, canonical and duplication problems, feeds. Cheap, fast, and frequently the reason a site with good content is invisible.

    robots.txt · crawler access2 blocked
    User-agent: GPTBotAllow
    User-agent: OAI-SearchBotAllow
    User-agent: PerplexityBotDisallow
    User-agent: Google-ExtendedDisallow

    Also checked: server-rendered body text, not text painted in by JavaScript.

  5. 05

    Weeks 2–3

    Define the prompt set

    Write the forty or so questions your buyers actually ask, in their words, across the whole funnel, including the comparison and alternatives questions you would rather they did not ask. That set becomes the yardstick everything else is judged against.

    Prompt set · 40 buyer questions5 engines
    best X for mid-market teamscited
    X vs competitor, which is betterabsent
    how much does X costcited
    alternatives to competitorabsent

    The same set is rerun on a schedule, so a change has a before and an after.

  6. 06

    Month 2+

    Read it and change course

    Rerun the set on a schedule, watch citation share and which sources moved, and put the effort where something responded. Same loop as every other channel I run, which is the point: this is not a separate practice bolted onto the side.

    Monthly readleading · business

    Leading

    Citation share, prompt coverage

    Business

    Branded search, assisted signups

    Shape of a report, not a forecast. Your baseline is whatever the first scan finds.

Corroboration is the one most providers leave out, and it is usually the one that moves. If that is the bulk of your gap, the community side has its own tooling.

Measurement

I built the tracker, so I am not reselling somebody else's dashboard.

Most providers in this category rent their measurement layer from a third-party platform and pass the invoice on. Mine is in-house. It runs your prompt set against every engine, records the full response, and keeps the sources so we can see which domains the answer was actually assembled from.

That last part is the useful bit. Knowing your LLM visibility score went up is mildly interesting. Knowing it went up because a community thread started ranking is something you can act on next week.

  • Your prompt set, rerun on a schedule, with the full answer stored
  • Share of voice against the competitors actually named in the answers
  • The exact source domains behind every citation
  • Included in the retainer, not billed as a software line
Look at the tracker
Dashboard – AI visibility
Scan complete

Visibility %

42%

▲ +11 pts

Total Mentions

318

▲ +64

Share of Voice

54%

▲ +8 pts

Avg Position

#3

▲ −0.6

Score by provider

Mention rate

ChatGPT
61%
Gemini
44%
Perplexity
52%
AI Overview
33%
AI Mode
28%
Sources – top cited
All sourcesDriving mentions
SourceProviderCitations
#1g2.com/categories/…ChatGPT41
#2reddit.com/r/…Perplexity33
#1yourbrand.com/blog/…Your domainGemini27
#3capterra.com/…AI Overview22
#2yourbrand.com/compareYour domainChatGPT18
The alternatives

Four ways to buy this. Three of them leave you holding the gap.

You have realistically got four options: hire an answer engine optimization agency, subscribe to a visibility tool and do it yourself, ask your existing search provider to add it on, or have one person own the whole thing. Here is the honest read on each.

busyless
AEO agency
Visibility tool
Your SEO provider
What you actually buy
One operator who owns the outcome across search, AI answers and the surfaces they cite
A packaged AI-visibility service, usually sold alongside the SEO retainer
A dashboard. The work is still yours to do
Rankings and traffic, with AI answers treated as a side report
Off-site sources
In scope, and usually where the movement is
Sometimes, often as a paid placement line item
Reported on, never acted on
Rarely in scope at all
Measurement
A fixed prompt set rerun on a schedule, plus branded search and assisted conversions
Vendor dashboard, resold
Whatever the tool measures
Sessions and rankings, which cannot see an answer
When it does not work
I tell you, and we move the budget
The retainer continues
The subscription renews
It gets reported as an algorithm update
Who does the thinking
The person you met on the call
A strategist, then juniors you never meet
You
An account manager relaying a specialist
Cost shape
$2,500 diagnosis, then from $5,000/mo to own it
Monthly retainer, usually annual
$100 to $1,000+/mo, plus your time
Monthly retainer

An AEO company is the right call if you need a large team pointed at a large site. An AEO consultant is right if you already have the people and only want the plan. This is for the middle, where you need both and cannot justify a hire for either.

Engagement

Find out what is wrong first. Then decide who fixes it.

Nobody should be quoting you a retainer before looking at your entity graph, your crawler access and what the engines currently say about your category. Most AEO services skip that step, because a diagnosis is where a scope usually gets smaller. So there are two doors, and the first one is small on purpose.

Start here

Discovery

$2,500one-off · two weeks

A baseline scan across every engine, a read of what is blocking retrieval, and a prioritised plan. You keep all of it whether or not we carry on, and it is specific enough for another provider to execute.

  • Baseline scan of your prompt set across every engine
  • Entity, structured data and crawler-access audit
  • Which sources the answers in your category currently use
  • A prioritised 90-day plan, sequenced by what unblocks what
Book the diagnosis

Most common

Ownership

From $5,000per month

All six workstreams, run and reported monthly, with the tracker included rather than billed on top. Month to month, because a provider who needs a year-long lock-in is telling you something about their confidence.

  • The six workstreams, owned end to end
  • The tracker and the monthly read, included
  • Judgement priced apart from production, so you can buy either
  • AI search optimization run inside the wider content strategy, not beside it
Talk it through

If you need volume rather than direction, production capacity is priced separately on the main pricing page. Buying leadership without buying volume is the point of splitting them.

FAQ

Questions, ahead of time.

  • What is answer engine optimization, in plain terms?+
    It is the work of becoming a source that AI answer engines retrieve and cite when somebody asks a question you should be the answer to. In practice that means making the facts about your business machine-readable, making your claims liftable in a single sentence, and being present on the third-party surfaces these systems lean on. Some people call it generative engine optimization. Same work, different label.
  • Is AEO actually different from SEO, or is it a rebrand?+
    It overlaps heavily and the overlap is the good news, because a well-structured, crawlable, genuinely useful page helps in both. The real difference is what you are competing for. Search gives you a slot in a list of ten. An answer engine picks a handful of sources and writes the summary itself, so you are competing to be quoted rather than to be ranked. That changes what you optimise and how you measure it.
  • Can you guarantee our brand gets cited in ChatGPT?+
    No, and I would treat any provider that does as a reason to leave the call. Nobody controls what a model outputs, answers vary between users and sessions, and the engines disagree with each other about sources. What can be done is to make you retrievable, consistent and corroborated, then measure citation share honestly over time and act on what moves.
  • How do you measure AEO when there is no rank and barely any referral data?+
    With a fixed set of about forty buyer questions, rerun across every engine on a schedule, so any change has a before and an after. That gives citation share, prompt coverage and which sources the engines pulled. Alongside it we watch branded search, direct traffic and assisted conversions. Attribution is partial here and I will say so rather than draw a clean line that does not exist.
  • How long does AEO take to show anything?+
    Retrieval and entity fixes can show inside a few weeks, because they remove a blocker rather than build an asset. Corroboration is slower and behaves like earned coverage always has, so give it a quarter. The first scan is the baseline, month two tells you which levers responded, and month three is the first read worth making a budget decision on.
  • Should we stop doing SEO and do AEO instead?+
    No. The engines are largely reading the same web, and several of them lean directly on a conventional search index to decide what to retrieve. Abandoning search to chase AI search visibility means removing the foundation the answers are built on. The right move is to keep the search work and change what the content is shaped like.
  • What does this AEO work cost here?+
    A two-week Discovery is $2,500 and ends with a diagnosis and a prioritised plan you keep whether or not we carry on. Ongoing ownership starts at $5,000 a month, which buys judgement rather than a deliverable count, and production capacity is priced separately on top if you need volume rather than direction.
  • Do you run the AEO work, or only advise on it?+
    Both, and they are priced apart on purpose. The retainer buys the thinking: the prompt set, the calls on what to fix and in what order, the editorial standard, and an honest read of the numbers. If you already have writers and a developer, you can buy the direction alone. If you do not, execution capacity is added on top.
Let's build

One call. Real plan, not a pitch.

30 minutes. We talk about what's already working, who owns content today, and whether a fractional Head of Content is actually the right move. If it isn't, I'll say so.

Direct calendar

Book a 30-min intro call

Loading scheduler…

Calendar busy?

Send a note instead.

One sentence on the bottleneck. I'll reply within 24h with a sharper next step.

Or send a note