Skip to content
Mavzena

How you work with the platform

Not a demo video: the real loop, module by module, the way you would run it week after week. See where you stand, fix what holds you back, and prove what worked.

The loop, start to finish

Three stages, five screens. Measure where you stand. Fix what holds you back: see on which prompts you fall behind, work through your next steps and make sure AI can read and trust you. Prove whether it worked, with a hypothesis set before you change anything.

  1. Stage 1 · Measure

    AI Visibility

    Start with where you actually stand

    Before changing anything, you run your brand's prompts against four AI engines (ChatGPT, Gemini, Claude and DeepSeek) in AI Visibility. It measures how often your brand gets mentioned, where it lands in the answer, and which named competitors keep showing up in your place: the baseline the rest of this loop builds on.

    Explore AI Visibility

  2. Stage 2 · Fix

    Opportunities

    Fix what holds you back, step by step

    Opportunities turns those answers into your next steps: which engines leave you out, who they recommend instead, and the organic work most likely to change it. Work that affects several prompts becomes a single step. Sometimes it is a comparison page, sometimes correcting a false claim the Truth Profile has found. Steps are listed in order of importance, each with its reason beside it.

    Explore Opportunities

  3. Stage 2 · Fix

    Organic Growth

    Make sure AI can read you, and trusts what it reads

    Organic Growth shows which sites AI answers in your category lean on and checks that AI crawlers can actually read your site. It also keeps the evidence you approve, so content generation cites it. In the Truth Profile you confirm the facts about your brand; false claims engines make about it are flagged against those facts, and you correct them from there.

    Explore Organic Growth

  4. Stage 3 · Prove

    Experiments

    On Growth and Agency plans

    Write down your hypothesis before you change anything

    You don't just publish and hope. Experiments attaches a hypothesis, the number you'll watch and a duration to the action before it runs, and keeps a record of past attempts so you don't quietly repeat one that already failed. It gives a verdict only after at least two new measurements made the same way as the starting one, because AI engines take time to notice a change.

    Explore Experiments

  5. Stage 3 · Prove

    Impact Proof

    On Growth and Agency plans

    Show whether it actually moved the number

    Every finished measurement is saved to Impact Proof on its own. The Impact Proof tab of Experiments compares the measurement before your change with the one after it and shows which metric moved, by how much, and how confident that comparison is: the evidence you need to keep going or stop, not a guess.

    Explore Impact Proof

What comes back, before you run one

Below is a small sample of the report the free page check gives you.

Check resultSample data

What AI crawlers can read

Technical score

66%

1 blocking · 1 recommended · 1 experimental

How AI engines describe this site

Not measured

This check measures whether AI can read your site. How often AI answers mention your brand, your visibility, is measured with a free account. Visibility was not measured on this page.

3 things to fix

In order of importance, blocking ones first.

  1. An AI crawler is blocked from reading this pageBlocking· As seen by GPTBot· -120 pts
  2. No Schema.org structured data was foundRecommended· -50 pts
  3. llms.txt is absentExperimental

See where your brand stands

Open a free account and we ask ChatGPT, Gemini and Claude 3 prompts about your market, so you see where you stand. No card needed.

Start free

Look around without signing up: an invented brand's 4-week AI visibility score, how it compares with competitors, what AI gets wrong about it, and the steps to take.