Being in the answer is not the same asbeing named in it.
Answer engine optimization for companies whose buyers now ask an assistant before they ask a supplier. Where the content in question is a help center or documentation set, the same discipline applies to structuring a knowledge base for answer extraction.
Your pages can be read, quoted and used to build the answer a buyer acts on, while your company is never mentioned. We measure that gap, and then we close it.
We are a small agency, not a platform. We use the tracking tools and we will tell you which one produced which number, but a dashboard reporting that your visibility fell does not fix anything. Working out which of several causes is real, and repairing it in the right order, is the job.
Three steps, and only the third one sells anything
1. Inclusion
Your page is in the set the model considered. Nobody outside the model can see this happen, and on its own it is worth nothing.
2. Citation
Your page is used as a source and appears in the reference list. Measurable, and the first thing most tools report.
3. Attribution
Your brand is named in the answer the buyer reads. This is the only step that sells anything, and it is the one almost nobody reports.
Most AI visibility reporting stops at step two, because citations are easy to count. Step three is what a buyer actually experiences, and it is where the work either paid off or did not.
The same sentence you already accept, two words changed
Search engines
Impressions are not the goal. Clicks are.
You were shown to someone. Whether they chose you is a different question.
AI search
Inclusions are not the goal. Citations are.
You were considered. Whether you were quoted, and whether your name survived into the answer, are two further questions again.
This is why AEO is sold and reported separately from on-page SEO. They act on different channels and are measured with different instruments. A client judging AI visibility work by search clicks will conclude it failed when it did not.
What measured AI visibility looks like
One client, tracked across the questions its buyers actually ask. It is the second most used source in its category, ahead of a national retailer, a major review publisher and two government agencies. Only a forum is used more.
Instrument and window: Searchable, 30 days to September 24, 2026, across ChatGPT, Perplexity and Google AI Overviews. Rank is by share of answers that draw on the source, not by raw citation count. One tool, one window, one client, and we are telling you which.
Distinct pages cited in AI answers, same category, 30 days
Read those last two numbers together
The same client is the second most used source in its category and is named in 5.8% of the answers it feeds.
It is cited constantly. It is credited almost never. A tool reporting citations would call that a triumph, and the buyer reading the answer still never learns who supplied it.
That is an entity and attribution problem, not a content problem, and no amount of extra writing fixes it. It is why brand mention rate sits beside citation count in everything we report, and why the work here is schema and identity before it is words.
The problem is usually not your content
A real diagnosis, on a live account
The client’s entire organization schema was invisible to the crawlers it was built for.
It had been delivered through a tag manager, so it existed only after JavaScript ran. Google renders JavaScript. GPTBot, ClaudeBot, PerplexityBot and CCBot generally do not.
The only statement of who this company was could not be read by a single one of the engines the AI visibility program existed to reach. The fix was to serve it from the server, not to write anything new. Verified afterwards by direct fetch and by a headless render.
This is the shape of most AEO work. Not more words, but making the words you already have reachable, parseable and attributable. It is unglamorous and it is where the movement comes from.
Why most AEO programs produce nothing
Six patterns, and we have found every one of them on live accounts.
Schema on pages AI cannot discover
Structured data on a page that takes a month to index does nothing. Competitor content is encountered first and becomes the baseline answer.
Optimizing content that is not machine readable
JavaScript-heavy builds, blocked resources and broken crawl paths mean the AI never reaches the content being optimized.
Skipping entity clarity
If your brand, products and categories are not recognizable entities, nothing you publish is attributable to you. This is the most common gap we find.
Treating it as a content project
Rewriting pages into question and answer format does not create AI visibility. Legibility and consistency across the whole estate do.
Working without a sequence
Random fixes do not compound. Each layer depends on the one beneath it, so investment accumulates without producing visibility.
Measuring the wrong thing
Tracking keyword rankings while ignoring citation share, indexation speed and brand mention rate leaves the real picture invisible to leadership.
The real cost is not the wasted budget. It is the compounding advantage a structurally stronger competitor builds while your fixes stay disconnected.
How we work
Diagnose
A full read of the estate: indexation, rendering, structured data, entity signals and current citation position. Bounded, paid, and it ends with a plan you could hand to someone else.
Sequence
A roadmap in dependency order, because the fixes only compound if they happen in the right sequence. Ranked by business impact, not by how easy they are.
Enable
Your team executes, with us alongside them. The capability stays in the building. We would rather you needed us less next year.
You can stop after the diagnostic, and some clients do. It ends with something you could act on without us, which is the point of buying a diagnosis rather than a retainer.
What you will see, and when
The prompt set and the baseline are recorded. Publish first and the baseline is gone permanently. There is no recovering it later.
Structural signals: indexation, schema validity, entity coverage. These move first because they are the ones we control.
Citation and attribution. Slower by nature, which is why a 28-day read alone would mislead you.
What the instruments cannot do, stated plainly. AI answers are non-deterministic, so a single query is a sample of one and we re-run any absence claim before it reaches you. Citation exports refresh only partially, so a period-on-period change is a floor rather than a measurement. Composite visibility scores are unstable enough that we will not quote them alone.
Who this is for, and who it is not
It fits mid-market B2B companies where digital visibility affects revenue, where an internal team can execute with direction, and where someone senior has noticed that buyers are arriving already informed by something nobody in the building wrote.
We will say no when:
- The SEO foundation is not there yet. If pages take thirty days to index, AEO is the wrong purchase this quarter and we will tell you so.
- Nothing can be baselined. Without a recorded prompt set before the work starts, no result afterwards is provable, and that baseline cannot be recovered later.
- You want a content subscription. Volume is not the lever here, and selling it as one would be dishonest.
- You want a guaranteed position in an AI answer. Nobody can sell that. Anyone who does is describing a screenshot, not a system.
Where this sits beside the rest
AEO is the middle layer. Search is the foundation beneath it, and generative engine optimization extends it into how models describe you when they are not citing anyone at all.
The on-page program shares the same pages and the same batches. The work overlaps. The measurement does not.
Two places it bites hardest: product catalogs, where assistants need structured product data before they can recommend anything, and local search, where assistants increasingly answer the question without showing a map at all.
Not sure which of these is your problem? That is what an SEO audit is for: it gathers the evidence, and the assessment tells you which finding is the cause and which are symptoms of it.
Questions we get asked
Q. What is answer engine optimization, in one sentence?
Making your business legible enough that AI assistants can find you, quote you accurately, and name you when they answer a buyer’s question. It covers indexation and rendering, structured data, entity clarity and content structure. It is not a writing exercise, although writing is part of it.
Q. Are you an AEO agency or an AEO platform?
An agency, and a deliberately small one. We are not a software company. We run the tracking platforms, including Searchable, where we are a certified partner, and we name the instrument behind every number we give you. What a platform gives you is a chart showing that something changed. What we do is diagnose which of several possible causes is real, fix it in dependency order, and prove the fix reached the crawlers that matter. If the chart alone is what you need, buy the tool and keep your money.
Q. How is AEO different from SEO?
SEO decides whether you appear in a list of links. AEO decides whether you appear inside an answer. They share a foundation, because a page that cannot be crawled, indexed or rendered is invisible to both, but they are measured with completely different instruments. SEO is judged on impressions, position and clicks. AEO is judged on citations, citation share and brand mention rate. Judging one by the other’s numbers is the most common way both get written off as failures.
Q. Do we need SEO before AEO?
Yes, and this is the prerequisite people most want to skip. If your pages take thirty days to index, or your content only exists after JavaScript runs, no amount of structured data will help, because the engines cannot reach what you are optimizing. The correct order is foundation first, then the AEO layer on top. We will tell you which one you actually need after the diagnostic, and sometimes the answer is that you are not ready to buy AEO yet.
Q. Is our FAQ schema still worth anything now that Google removed FAQ rich results?
Yes, and this is a good example of why. Google withdrew FAQ rich results for most sites, so the original payoff is gone. The markup kept its value anyway, because answer engines read it. We have clients whose structured question and answer estates were built well before AEO had a name, and those pages are now among the most cited in their category. The lesson is not that we predicted it. It is that structural discipline survives a change in who is reading.
Q. How do you measure whether it is working?
Citations, citation share on the questions that matter, the number of distinct pages being cited, and brand mention rate. The last one matters most and is reported least: a page can be cited while your name never appears in the answer. We record the prompt set and the baseline before any work starts, because after publication the baseline cannot be recovered.
Q. Which AI platforms does this target?
The ones your buyers actually use: Google AI Overviews and AI Mode, ChatGPT, Perplexity, Gemini and Copilot. The underlying work is largely the same for all of them, because they are all solving the same problem of deciding which sources to trust. Where they differ is in measurement, and no single tool sees all of them well, so we say which instrument produced which number.
Q. How long before we see anything?
Structural signals such as indexation, schema validity and entity coverage move within about 28 days because they are the part we control directly. Citation and attribution are slower, so we set a 60-day read alongside the 28-day one. Anyone promising AI visibility in two weeks is describing a single lucky query rather than a measurement.
Q. Can we do this ourselves?
Often, yes, and that is the intended outcome. Most of this work is executed by your own team. What is hard to acquire internally is the sequence and the judgment about which gap matters first, because the fixes only compound in dependency order. We aim to leave the capability behind rather than the dependency.
Q. What can these tools not tell us?
More than the vendors admit. AI answers are non-deterministic, so a single query is a sample of one, not a measurement, and we re-run any claim that something is absent before it reaches you. Citation exports refresh only partially, so a period-on-period change is a floor rather than a number. Composite visibility scores are unstable enough that we will not quote them on their own. We report the limit next to the figure.
Start with a conversation
Thirty minutes on where you currently stand in AI answers and whether the gap is structural or editorial. We can usually tell early. If the honest answer is that you need search foundations first, we will say that rather than sell you the more fashionable service.