Be the answer AI gives. An AI search agency that shows its working.

Answer engine and generative engine optimization for companies whose buyers now ask a model before they ask a search box. Crawler access, entity resolution, writing built to be quoted, and original data your competitors do not have — with the method published, the assumptions stated and the corrections logged in public.

Work With Us How AEO Works
7
AI crawlers audited,
engine by engine
51
States priced live
on a client build
3
Corrections published
on that client site
Scroll
~1,000Organic LinkedIn followers from an hour a week of voice notesB2B sales leader · 3 months
100,000+Platform-reported views, starting from no public presence at allManufacturing software founder
51States priced from a live government API on a client SEO buildheatpumpmath.com
1 hrA week of your time. That is the whole input we ask forHow that becomes a week of output

Four engagements, written up with the timeframe on every figure and a stated list of what we are not claiming. All case studies.

Your buyers now ask a model first. It is answering with someone else’s name.

Somewhere this week a buyer asked ChatGPT, Perplexity or Google’s AI Overview who the credible option in your category is. It returned three or four names, cited a handful of sources and moved on. If you were not in that answer you were not in the consideration set — and unlike a search result you were beaten by, there is no page two to be on.

The uncomfortable part is that this is rarely decided by how good you are. It is decided by whether the engines can fetch your pages at all, whether they can resolve you to one identifiable company, and whether anyone independent has said the same thing about you twice. Most firms fail the first test without knowing it, because a security product or CDN setting nobody chose is returning 403 to the crawlers that feed those answers.

That is the work: reachable, resolvable, worth quoting — then measured honestly enough that you can tell whether it moved.

An experienced professional at their office desk

The Entity Stack

A model cannot recommend
a company it cannot resolve.

Answer engines do not rank pages, they describe entities. Five layers, built in this order, because each one is wasted without the one beneath it — and the first is the one most firms fail without ever being told.

An experienced professional at their office desk
Depth 01 · The Floor

Crawler Access

Whether GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Google-Extended can actually fetch your pages, confirmed from server logs rather than guessed from robots.txt. Blocked at the edge by a setting nobody chose is the most common failure we find, and nothing above it matters until it is fixed.

Depth 02 · The Identity

Structured Data & sameAs

Organization and Person markup describing something real, one canonical identity page, and a sameAs graph tying your profiles together. This is what turns a string that happens to be your name into an entity a machine can tell apart from the three other firms sharing it.

Depth 03 · The Record

Wikidata

The machine-readable record of the company: what it does, where, since when, and which identifiers it maps to. A far lower bar than Wikipedia and a far more direct signal, where the company genuinely qualifies. Where it does not, we say so rather than create one that gets deleted.

Depth 04 · The Panel

Google Knowledge Panel

Google’s public read on who you are, generated once its knowledge graph is confident. There is no form and no paid route, so every lever is indirect: it is the output of the three layers beneath it, which is exactly why it cannot be bought.

Depth 05 · The Answer

Cited in AI Answers

ChatGPT, Perplexity, Gemini and AI Overviews composing a reply and attributing it to you. Measured against a fixed prompt set with the sample size reported, not screenshotted once and called a trend.

Then the loop closes. Fresh articles become fresh citations, fresh citations strengthen the Wikipedia page, and the whole system compounds while you keep doing the work you are actually known for.

The Technical Floor

Most firms fail here.
Almost none of them know it.

Before anything gets written, three things have to be true: an engine can fetch your pages, it can render them, and it can tell which company you are. Fail any one and every piece of content published afterwards is invisible to the systems you are trying to reach.

This is the least glamorous work we do and the fastest to show a result. It is also the part a content retainer never touches, which is why so many of them run for a year without moving anything.

Crawler Reachability

Confirmed per engine from your server logs, not inferred from robots.txt. A permissive file proves nothing when a CDN bot manager or security product is quietly returning 403 to ClaudeBot. This is the single most common fault we find, and usually the fastest to fix.

Rendering Without JavaScript

Whether your pages contain their content in the HTML, or assemble it in the browser afterwards. Retrieval systems vary in how much JavaScript they will execute, and a page whose substance only appears after hydration is a page that may be read as empty.

Entity Resolution

Structured data describing a real organization, one canonical identity, a sameAs graph, and consistent details wherever you appear. Without it a model has a string, not a company — and strings get confused with competitors who share them.
Realistic Timeline
Access faults, once found, resolve on the crawler’s own schedule — usually days to weeks, and the fastest result in the engagement. Entity resolution shows up as models and knowledge graphs refresh. Earned position takes months and depends partly on third parties, which is why nobody honest will give you a date for it.

The Company Engagement

Six things, in this order.
Each one wasted without the last.

Month one · Diagnostic

The Access Audit

Seven AI crawlers checked against your server logs, engine by engine, plus a render check. We find out whether the systems you are trying to reach can read you at all before anyone writes a word.
01
Ongoing · Technical

Entity Resolution

Organization and Person structured data, one canonical identity, a sameAs graph, corrected directory records, and a Wikidata item where you genuinely qualify. This is what makes you a company rather than a string.
02
Ongoing · Editorial

Extractable Rewriting

Your existing pages restructured so each opens with a complete, specific, checkable answer. Names, dates and numbers survive extraction; positioning language does not. Usually a rewrite, not a publishing schedule.
03
Where it exists · High value

Original Data

The most defensible page in any category holds a number nobody else has. Sometimes a public API nobody in your sector has pulled, sometimes something measured directly. It is the one thing a competitor cannot copy the week after they notice you.
04
Ongoing · The slow half

Corroboration

Getting independent sources to state the same facts the same way. Published research repeatedly finds citations skewing toward earned third-party coverage over a brand’s own domain, which makes this the half that actually moves the answer.
05
Monthly · Reported

Monitoring & Reporting

A frozen, versioned prompt set re-run on schedule, with presence rate, citation share and the competitors named instead of you — every figure published with its sample size. The formulas are public.
06

The Diagnostic

This is what we show you on the first call.

We run the questions your buyers actually ask across the major engines, clean, several times each, and record what comes back. Illustrative below — the real version names your competitors and cites their sources.

ChatGPT
“Who are the best [category] firms?”
Based on available information, the most frequently referenced firms in this space are Competitor A, Competitor B and Competitor C…
cited: trade-publication.com cited: competitor-a.com
You: not named
Runs: 5 of 5
Perplexity
“Is [your company] any good?”
I could not find detailed information about this company. There is a firm with a similar name operating in a different sector…
entity: ambiguous crawler: 403
Cause: unreachable
Fix: weeks
Google AI Overview
“[Category] pricing”
Pricing in this category typically ranges… according to Competitor B, which publishes its rates publicly.
cited: competitor-b.com
Why them: they published
You: no page

None of those three is a content problem. The first is corroboration, the second is crawler access and a confused entity, the third is a page that does not exist. That is the order we work in, and it is why the audit comes before anything gets written.

The Founder Track

The second offer, for the person
rather than the company.

Some buyers are a company. Some are a founder or an experienced professional whose reputation stops at the edge of their own network. The search work above applies to both, but the individual track adds distribution — publishing under your own name until the industry treats you as the person who knows. Same $2,000 a month, same three month minimum. Below is what that track runs.

01
Every Week

LinkedIn, Consistently

Three to five posts a week in your voice, built from a weekly session with you. Not motivational filler. Actual positions on your subject, the kind colleagues argue with and journalists screenshot.
3-5 Posts / WeekYour VoiceYou Approve
02
Every Week

AI Avatar Video

Video built from your own likeness and voice, so you publish on camera without filming every piece. Reviewed and approved by you before anything goes live, and it is you, not a synthetic persona.
Your LikenessShort FormMulti-Platform
03
Ongoing

Podcasts, Arranged

We find and book the shows your buyers already listen to, and prepare you for them. Each appearance is a durable citation with show notes, a transcript and a link, working long after the episode drops.
PlacementPrepShow Notes
04
Monthly

Long-Form On Your Own Site

Substantial articles under your name, on a domain you own. This is the asset nothing can take away, and the one search engines and AI tools resolve you against when someone looks you up.
Your DomainQuotableIndexed
05
Ongoing

SEO & AI Search

Technical SEO on the site you own, plus the entity work that lets Google and the answer engines resolve who you are. Where your field has data nobody has published, we publish it, with the sources named and the method shown. How we run it.
TechnicalOriginal DataAEO
06
Ongoing

PR & Earned Coverage

Pitching you to journalists and publications as a source worth quoting on your subject. This is the layer everything later depends on, because coverage written by other people is the only kind that counts.
Media PitchingCommentaryFeatures
07
Ongoing

Distribution Across Channels

The same thinking, reshaped for every surface your audience actually uses, because trust is built across places rather than in one. Being seen once is not being known.
Multi-ChannelRepurposingCompounding

Sustained long enough, that publishing produces the independent coverage every later layer depends on — and for a minority of people it eventually reaches the point where an encyclopedia entry becomes defensible. That is the end of the chain, not the start of it, and for most clients it is untouched in the first year. The honest timeline for each stage.

The Process

Audit first.
Then fix what the audit found.

1
Step One

The Diagnostic

Crawler access checked per engine against your logs, a render check, an entity audit, and a prompt set built from the questions your buyers actually ask — run clean, several times each. You leave the first call knowing which of the three failure modes you have.
Niche MappingAvatar BuildSite Live
2
Step Two

The Floor

Access faults fixed, structured data and sameAs in place, directory and database records corrected, Wikidata where you qualify. Unglamorous, fastest to show a result, and the part a content retainer never reaches.
Podcasts BookedPR LiveWikidata Live
3
Step Three

Quotable, Then Corroborated

The pages that matter rewritten to open with a complete, checkable answer. Original data published where your sector has data nobody has pulled. Then the slow half: getting independent sources to state the same facts the same way.
Inbound StartsPress QuotesInvitations
4
Step Four

Re-measured, Honestly

The same frozen prompt set re-run on schedule. Presence rate, citation share, and who is still being named instead of you — each figure reported with its sample size, and movement inside the noise floor called what it is.
Panel ActiveAI Cites YouWikipedia Possible

Fit

This works for a specific company.
It is slow and frustrating for everyone else.

Three or four calls a month end with us saying this is the wrong purchase. That is cheaper for everyone before the invoice than after it, so here is the honest version in advance.

A good fit

Your buyers research before they buy

  • Your category gets asked about. People put questions to a model before they shortlist, which is true of most considered B2B and professional-services purchases and untrue of impulse and walk-in trade.
  • You have something checkable. Data, measurements, operating detail, a public dataset nobody in your sector has pulled. This is the fastest route to pages a competitor cannot copy.
  • You will let us touch the site. The technical floor and the existing pages, not just a content calendar bolted on the side.
  • Someone can sign off. A person who can approve a claim and say whether it is substantiated, which matters more in regulated categories.
  • You are buying position over a year. Not sessions this quarter.

Not yet, or not us

You want the answer engine to flatter you

  • You need attributable pipeline this quarter. AI referral traffic is around 1% of web traffic. Buy demand generation; it works on that timescale and this does not.
  • You want a guaranteed mention, ranking or panel. No vendor controls any of them, and there is no ranking inside an answer engine to hold.
  • You want volume. A fixed number of posts a month against a keyword list. A content agency will do that better and cheaper.
  • Nothing on the site can change. If the brief is new content only and the existing pages and technical setup are off limits, we would be charging you to decorate.
  • You need a large team or fast turnaround. A real constraint at our size, and not one money solves. Several firms we assessed are a better fit.

If the second column describes you, say so on the call and we will tell you what would actually help, including when that is not us. What the work consists of · what it costs.

Investment

One price, published.
Nobody else in this field publishes one.

Company and founder engagements alike
One Engagement
Crawler access · Entity resolution · Extractable pages · Original data · Corroboration · Measurement
$2,000
per month
No tiers, no setup fee
Three months,
then month
to month.
The Access Audit
Seven AI crawlers checked against your server logs engine by engine — GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended — plus a render check. Run first, because nothing above it matters until it passes.
Entity Resolution
Organization and Person structured data, one canonical identity, a sameAs graph, corrected directory and database records, and a Wikidata item where you genuinely qualify. Where you do not, we say so rather than create one that gets deleted.
Extractable Rewriting
Your existing pages restructured so each opens with a complete, specific, checkable answer. Usually a rewrite of what you already have rather than a publishing schedule bolted on beside it.
Original Data, Where It Exists
A public dataset nobody in your sector has pulled, or something measured directly. The one part of the work a competitor cannot replicate the week after they notice you.
Corroboration
Getting independent sources to state the same facts the same way. Published research repeatedly finds AI citations skewing toward earned third-party coverage over a brand’s own domain, which makes this the slow half that actually moves the answer.
Monitoring and Reporting
A frozen, versioned prompt set re-run on schedule: presence rate, citation share, and the competitors named instead of you, each figure reported with its sample size. The formulas are published so you can reproduce them.
Founder Track: LinkedIn and Video
For individual clients rather than companies. Three to five posts a week in your voice from a weekly session, plus video built from your own likeness, reviewed and approved by you before anything goes live.
Founder Track: Long-Form and Distribution
Substantial articles under your name on a domain you own, podcast placements, PR outreach, and scheduling across the channels your audience actually uses.
The Record Layer, Eventually
Google knowledge panel groundwork, and Wikipedia only where independent coverage genuinely meets the bar — which for most clients is not year one. Disclosed as the Wikimedia Terms of Use require, and submitted through Articles for Creation.
Book the Diagnostic →

3-month minimum, then month to month · few clients at a time · what this does not buy

Questions

What companies ask us.

Why isn’t our company showing up in ChatGPT or AI Overviews?

Usually one of three things, and almost never the one people expect. Either the engines cannot fetch your pages — commonly a CDN bot manager or security product returning 403 to AI crawlers, which is invisible in robots.txt — or they cannot resolve you to a specific company and are confusing you with a similarly named one, or your pages contain positioning language rather than specific checkable facts a model can quote. The audit tells you which, in the first week.

Is AEO actually worth buying yet, honestly?

Depends entirely what you are buying it for. Published figures put AI referral traffic at roughly 1% of total web traffic, so if the goal is sessions this quarter the honest answer is no, and we will say so on the call. If the goal is being the name that comes back when a buyer asks a model who the credible option is, the work compounds and the category is far less contested than organic search. We would rather lose the sale than have you buy it on a traffic forecast.

Can you guarantee a model will mention us?

No, and neither can anyone else. Answers vary between runs, models change without notice, and no vendor has a lever that inserts a name into a generated answer. There is also no ranking position inside an answer engine to hold, so any figure presented as one was constructed. What can be committed to is the access, the structure and the corroboration that make citation possible, plus honest measurement of whether it is happening.

How is this different from our current SEO agency?

A good deal of it is ordinary good SEO and we say so rather than dressing it up. The genuine differences are that crawler access for AI engines is decided separately from Googlebot and is frequently broken, that pages get restructured for passage extraction rather than page ranking, that entity resolution is a deliverable rather than a nice-to-have, and that measurement runs against a frozen prompt set instead of a rank tracker.

What do you actually report each month?

Presence rate across a fixed prompt set, citation share, the competitors named instead of you, crawler reachability per engine with the date checked, and the pages changed with what the passage said before. Every figure carries its sample size, and movement inside the noise floor gets called noise. Every formula is published, including what we refuse to report.

What does it cost, and what is the commitment?

$2,000 a month with a three month minimum, then month to month. One price for company and founder engagements alike — no tiers, no setup fee, no per-article rate, and the first call is free. None of the five agencies we assessed in this field publishes a price at all. Ours, and what it does not buy.

Ask a model about your category today.
Then look at whose name comes back.

That answer is already being given to your buyers, several times a day, whether or not anyone at your company has read it. The first call is a diagnostic: we run the prompts, show you who is named instead of you, check whether the engines can reach your site at all, and tell you honestly whether this is worth buying in your category yet.

Book the Diagnostic →

Free, thirty minutes, and it often ends in a no  ·  few clients at a time