Home Services AI Search Optimization
AEOGEO

People can find your site.
Can the AI?

ChatGPT, Google, Claude, Perplexity, Copilot and Grok answer your customers' questions — and they name specific companies. The only question is whether yours is one of them.

Free check — once per website. It shows where you stand today; tracking whether things improve needs a repeated measurement, which is part of the paid work. We reply within one business day and email the check itself within three.

Why now

The new search
works differently

When your customer asks ChatGPT "who does [your service]" — what do they hear back? For a little while longer, that answer is still changeable.

It's SEO in 2009 again: the companies that did the work then still get free traffic. The ones who waited missed it.

AI search only works with what it already knows. Without this work, you are simply not on the list.

AI already recommends someone.
Just not you.

Ask ChatGPT "who does [your service]" and it names real companies. Here's what we check to see why they got there — and you didn't.

  • 01Whether their pages carry structured markup that AI can lift facts from
  • 02Whether they have a page built for the exact query your customer is asking
  • 03Whether they have reviews and third-party mentions across the sources AI reads
  • 04Whether they show up on the platforms and directories AI models pull from
  • 05Where the AI's answer is actually sourced from — and whether that source can be matched

This is the first thing we run in the audit: not just how AI sees you, but who it names instead — and exactly what they have that you don't yet.

What the optimization covers

Seven steps, from audit to a report and a playbook for your team. Transparent: you see what was done and what has to be maintained.

Step 1 · Audit
How AI sees you today
We check what ChatGPT, Google, Claude, Perplexity, Copilot and Grok find about you, what they ignore, and which facts are missing.
Step 2 · Competitors
Who AI names instead of you, and why
We identify the companies AI recommends for your queries and reverse-engineer what got them there: markup, content, reviews, presence.
Step 3 · Technical files
llms.txt, robots.txt, sitemap.xml
We create and configure llms.txt — the file that tells AI crawlers who you are and what you do. We review and update robots.txt and sitemap.xml.
Step 4 · AI-ready content
The structures AI lifts answers from
We reformat key pages: FAQ blocks, hard facts and numbers, question-and-answer structures — exactly what models quote when they answer a user.
Step 5 · Company and service pages
The factual format AI prefers
We rewrite them to be concrete: who you are, what you do, for whom, with what results. AI rewards specifics, so we give it specifics.
Step 6 · Analytics
Tracking traffic from AI assistants
ChatGPT and Perplexity referrals often show up as "direct traffic" in analytics. We set up UTM tagging and AI-referral segments so this traffic becomes visible.
Step 7 · Report and playbook
So you can maintain it yourself
We document everything that was done and hand your team a playbook: what to update when things change, how to write new pages and add services so AI picks them up.

Your involvement
is minimal

We take care of the rest.

/ 01
Access to the site, or a contact for your developer
/ 02
Answers to a handful of questions about the company and its services
/ 03
Your product expert on standby — to confirm accuracy on specialized questions
/ 04
Sign-off on the final result

How we measure

Visibility in AI answers is not a single lucky screenshot. Here is the methodology behind every number we report.

6
AI systems
ChatGPT, Google (AI Overviews, AI Mode and Gemini), Claude, Perplexity, Microsoft Copilot and Grok. We count vendors, not screens.
50–80
Queries
Brand, category, commercial and comparison phrasing, in natural language.
5
Runs per query
AI answers vary between runs. We track how often you appear — "named in 2 answers out of 5" — not a single lucky screenshot.
2
Before / after
The same methodology both times, so the numbers are actually comparable.
1,500–2,400
Answers per measurement
A typical project: 50–80 queries × 6 systems × 5 runs. That is one measurement. There are two — a baseline before the work starts and a follow-up after, run with the same methodology so the numbers are comparable.
A number you can trust
Models answer the same question differently. We do not hand you one pretty percentage from a single run. We measure how often you appear across repeats — and show how stable that number is. Otherwise the work plan is built on a good day or a bad day.
More than one chat window
A chat answer is only one layer. We also cross-check with independent signals: what the platforms themselves report about your visibility, and whether AI systems can actually reach your pages. Often the chat "knows" you while access to the site is blocked — or the reverse. A ChatGPT screenshot alone will not catch that.

The system set follows your buyers. The six above are the base set — they cover where most US buyers actually ask. If your audience lives somewhere else, we swap one in at no extra cost; the methodology does not change. Need more than six systems covered, or a region with its own assistants? We will do that too — quoted separately.

Case studies are under NDA. But you can verify us without them.

We apply what we sell to ourselves — our llms.txt is public, see it here.

Ask ChatGPT Ask Claude Two of the six systems we measure — the same question works in any of them.

Honest about results

No inflated promises. Here is exactly what the work does and does not cover.

We guarantee
  • A baseline measurement before work starts
  • All work that affects visibility, delivered in full
  • A follow-up measurement, using the same methodology
  • Full reporting on what was done and what changed
We don't guarantee
  • A specific ranking or position inside AI answers
  • A specific timeline for results to appear
  • That competitors won't do the same thing

One thing is certain: a model can't recommend a company it doesn't know about. Right now it doesn't know about you — that's what we're changing.

Who this is not for

To save both sides time:

This is probably not the right fit if
  • You need to "be in ChatGPT" within one month — the market does not work that way
  • You will not update content or company facts — technical files alone barely move visibility
  • You compete only on price and have nothing distinctive for a model to say
  • You want a pretty PDF with advice and no implementation — we sell work plus a remeasurement, not a folder of tips

If that is not you — start with the free check.

Pricing

Start free, then go as deep as the site needs. We give an exact number after a short call, once the scope is clear.

Express check
Free

10 queries across 3 AI systems — you pick the three; ChatGPT, Claude and Perplexity by default. A snapshot of where you stand at the moment we check: what AI says about you and who it names instead.

Once per site

Full audit
$1,500one-time

50–80 queries, 6 systems, 5 runs each, competitor analysis, a prioritized task register, and a 90-day growth roadmap.

Credited toward implementation if you book both

Implementation
$3,500one-time · 2–4 weeks

The foundation, delivered end to end: technical files and markup, AI-ready content on key pages, company data unified across sources, analytics tracking. Closing measurement included.

A complete piece of work — after it, AI can find and understand you

Ongoing work
from $3,500/mo

Closing new gaps every month: content for queries where AI still doesn't find you, review management, listings, external mentions, and a monthly remeasurement.

Optional — start any time, cancel any time

Pricing covers our work. Paid placements, media budgets and third-party fees are billed separately, at cost, and always agreed with you first. Prices are indicative and depend on project scope; this is not a public offer. Running ads inside ChatGPT is a separate, paid-placement service — see AI4Live ChatGPT Ads Management.

Where implementation ends
and ongoing work begins

Implementation is a complete piece of work. We fix what's broken and build the foundation: technical access for AI crawlers, markup, llms.txt, key pages rewritten in the format models quote, company facts unified across sources, analytics in place. After it, AI can find you and understands what you do.

Ongoing work is different. It isn't finishing the same job — it's closing new gaps, month after month: content for queries where AI still names competitors instead of you, reviews, listings, mentions in the sources AI actually cites. This never "completes", because your competitors don't stop either.

You can stop after implementation. Most clients don't, because the second part is where the visibility compounds.

Ongoing work, broken down

from $3,500/mo. Three levels, depending on how much ground you want covered every month.

Essential
$3,500/mo

Monthly remeasurement across all systems, competitor gap analysis, technical maintenance, 1 content piece per month.

Growth
$5,500/mo

Everything in Essential, plus 3 content pieces, review management, listings and profile updates across the sources AI cites.

Full
$8,500/mo

Everything in Growth, plus external placements, ongoing mentions and PR work, priority support.

Pricing covers our work. Paid placements, media budgets and third-party fees are billed separately, at cost, and always agreed with you first.

Frequent questions

Briefly: how AI search optimization works and what to expect.

SEO is about positions in Google, where the user gets a list of links. AEO — Answer Engine Optimization — is about the AI naming your company inside the answer itself. Users increasingly never click a link at all: they ask ChatGPT and act on what it says. If you are not in there, you do not exist.

A file in the root of your site that AI crawlers read to understand who you are and what you do. Think robots.txt, but for language models.

The base implementation is one-time. After that it needs upkeep whenever the site changes or you add services. We hand over a playbook so your team can do it, or we run the upkeep ourselves every month.

Yes. A landing page or a large corporate portal — the approach is the same. Only the volume of work and the price differ.

AI crawlers index sites at different speeds. Changes usually start showing up in answers within 4–8 weeks.

No — and anyone who guarantees it should not be trusted. The algorithms are opaque and they change. We do everything technically possible to put you in front of AI answer engines. Without this work there is no chance at all.

The base set is six: ChatGPT, Google (AI Overviews, AI Mode and Gemini), Claude, Perplexity, Microsoft Copilot and Grok. That is where your buyers actually ask, so those are included by default. We count vendors, not screens — Google's three surfaces answer differently and we measure all three, but that is one system on the list, not three.

The list is not fixed. Swap any one of them for something else — Meta AI, DeepSeek, an industry assistant, your own internal model — at no extra cost. The methodology does not change. If you need more than six systems covered, we will do that too; it is quoted separately.

The free express check covers exactly three systems. That is a limit on volume, not on choice — you pick which three. We default to ChatGPT, Claude and Perplexity.

If your audience has a quirk — buyers who live inside a vertical tool, or a corporate assistant of your own — tell us and we will work it out. How we choose the six →

Model coverage stopped being a differentiator in early 2026. What we do differently is how we measure and what we do with the numbers:

  • Stability, not a vanity score. Repeated runs so the action list is not built on a fluke answer.
  • More than one layer of measurement. Not only "what the chat said" — cross-checked with independent visibility signals and whether AI can actually reach the site.
  • Accuracy of what is said about you. Not only "mentioned / not mentioned", but what is said — outdated prices, wrong location, retired services.
  • Competitors by substance, not a name list. Why they get recommended instead of you — and in what order to close the gaps.
  • Before/after with the same method. No guarantee of a rank inside the answer — a guarantee of a transparent comparison.

Because the same question gets different answers on different runs. One screenshot is a coin flip. We run each query multiple times and report frequency — "named in 2 out of 5 answers". Otherwise a good day looks like a win and a bad day looks like failure.

You can ask ChatGPT or Claude about your company and see what comes back. Understanding why a competitor gets named instead, and closing that gap, is the part that needs dedicated work.

That is possible in a competitive niche. It is exactly why we measure before and after with the same methodology and adjust the plan based on what the numbers show.

Referrals from ChatGPT and Perplexity often show up as "direct traffic" in analytics. We set up UTM tagging and AI-referral segments so this traffic becomes a visible channel.

Each month: a check of who AI names instead of you and why, new content for the queries where you are missing, review management, updated company facts across sources, technical file maintenance, and a follow-up measurement.

Case studies are under NDA. You can verify us without them: ask ChatGPT and Claude what they know about AI4Live. We apply what we sell to ourselves — including a public llms.txt.

Fill in the form above with your company name and website. One report per site. We reply within one business day and email the check within three. We also flag where the model describes you incorrectly — that is usually what makes the follow-up call worth having.

Optimization is free organic visibility built over weeks. ChatGPT ads are paid placement through an ad auction, with a separate budget you control directly — see ChatGPT Ads Management.