avowdFree check

Generative engine optimization

Vouched for
by the machines.

Your customers stopped searching — they ask an AI. We measure what nine engines actually say about you, including the Chinese ones nobody else can reach, then make your name the answer.

No credit card · 30 seconds · 9 engines

Right now, in someone’s chat window

The shortlist forms without you.

Your brand isn’t in that answer. The customer never sees you, never bounces off your site, never shows up in your analytics. The loss is invisible — that’s what makes it expensive.

Why it happens

Search moved into the chatbox.

Nobody announced it. There was no migration, no redirect — buyers simply started asking a machine instead of typing into a box, and the machine started answering with three names.

01

Buyers ask AI first

Before they open a website or read a review, people ask an assistant who to choose. That reply becomes the shortlist they act on.

02

It names your competitors

When a model doesn’t know or trust you well enough, it doesn’t hedge — it confidently recommends someone else, by name, in your category.

03

You never see it happen

No click, no bounce, no line in your analytics. The customer you lost looks exactly like a customer who never existed.

Coverage

Nine engines. Two worlds.

Every competitor measures the same five Western models. We measure those — plus the Chinese engines that are closed consumer apps with no public API, where a huge share of buyers actually ask.

ChatGPT
West
Perplexity
West
Claude
West
Grok
West
Gemini
West
Doubao 豆包
China
Qwen 通义
China
DeepSeek
China
Kimi
China

No guesswork, no black hat. Every number comes from a live API call we can show you. We don’t buy reviews and we don’t seed fake threads — everything we build for you is something you’d be happy to show a customer.

How it works

Measure. Fix. Hold the ground.

1

Measure what’s true today

We ask every engine the questions your buyers actually type, several times over, and count what comes back. That’s your baseline — a number, not a feeling.

2

Fix what the models read

Structured data, answer-first pages, llms.txt, and citations on the third-party sources the engines already trust. In English, in Chinese, or both.

3

Defend the position

Models retrain and rivals chase. We re-measure monthly, replace citations that decay, and tell you plainly if you slipped — and why.

What you actually get

Everything measured, nothing invented.

Each of these is a real surface in your report — the sample figures below are illustrative.

One number

A score you can put in front of a board.

Mention rate, share of voice and citation rate compressed into a single 0–100 read, with the three components broken out so nobody has to take it on faith.

Example · visibility score

38needs work
Mention42%
Share of voice29%
Citations18%

In their words

How each AI describes you, verbatim.

The fastest way to find out a model thinks you sell something you don’t. We ask all nine to describe your business in a line and hand you the raw answers.

Example · engine descriptions

ChatGPTA Canadian payroll and HR platform for small businesses.
PerplexityA payments company serving retail merchants.
豆包(no reliable information about this brand)

The lever

Who gets cited instead of you.

Every domain the engines lean on when answering your questions, ranked. The ones marked as openings are exactly where the next month of work goes.

Example · cited sources

wikipedia.orgopening
reddit.comopening
competitor-a.comrival
g2.comopening
zhihu.comopening

The plan

A checklist, ordered by what moves first.

Quick wins on your own site, then the slower off-site work. The same list we execute from — so you can hand it to your own team instead of us, if you’d rather.

Example · action plan

  • 01Publish an answer-first FAQ on your three buying questions
  • 02Add Organization + Product structured data
  • 03Serve an llms.txt describing what you actually do
  • 04Earn a citation on the two sources the engines already trust
  • 05Establish the Chinese-language entry the models can’t find

What changes

Before and after, honestly.

Without us

  • You don’t know whether AI is recommending your rivals
  • Customers vanish and nothing appears in your analytics
  • No idea what the models think you even do
  • The Chinese engines are a complete blind spot
  • Checking by hand, occasionally, with no consistency

With us

  • A real number every month, for every engine
  • You see exactly which sources get cited instead of you
  • You read the models’ own words about your brand
  • Both halves of the world, measured the same way
  • Work executed, not just a dashboard to log into

Start with the audit. Continue only if it earned it.

No lock-in and no retainer you can’t leave. The audit is yours to keep either way — plan included, whether or not you carry on.

01 — the audit
$750 one-time

Nine-engine baseline, scored report, how each AI describes you, who gets cited instead of you, and a prioritized fix plan.

02 — the defence
$600 / month

We execute the plan, then hold the ground: monthly re-measurement, off-site citations, content refresh, alerts when an engine turns. Cancel anytime.

For the skeptics

The questions we’d ask too.

Isn’t this just SEO with a new name?

No. SEO wins a position on a page of ten blue links. This is about being named inside a generated answer, where there is no page and usually no list — often just one or two brands. Different signals, different work, different measurement.

You did the work once — why keep paying?

Fair, and the honest answer is that the technical foundation is one-time. What isn’t one-time: models retrain, sources get updated, and your competitors are doing this too. The monthly fee buys measurement and defence, not repeated setup — and you can stop it the month you decide it isn’t earning.

Can you guarantee AI will recommend me?

No, and anyone who guarantees it is lying to you. We guarantee honest measurement, real work on the signals that demonstrably move, and that you see the number either way — up or down.

Do you buy reviews or spam forums?

Never. That’s the black-hat corner of this industry and it gets brands burned when platforms clean house. Everything we build is something you’d be comfortable showing a customer.

Why can you measure Chinese engines when others can’t?

Because we built for them from the start and we work in both languages. Doubao, Qwen, DeepSeek and Kimi need separate access, separate prompt sets and someone who can read the answers. Most platforms simply skip that half of the map.

Find out what the machines say about you.

One domain, thirty seconds, no card. You’ll see whether nine engines know you exist — and what they tell your customers when they don’t.

No credit card · 30 seconds · 9 engines