How the audit works
I run 20 real buyer questions through three AI engines and log every answer and every source they cite. You get a PDF showing where your company is missing, who takes those answers instead, and which pages to fix first
The questions
Not keywords. The questions someone actually types before they buy.
Twenty per category, covering the three shapes a buyer question takes:
the broad one, "best X software 2026"
the price one, "cheapest X for a small team"
the alternatives one, "best alternatives to [the leader in your category]"
The rest cover specific needs — a feature, a company size, an integration, a platform.
Comprehensive audits with tracking
Veza Digital — 4,500 USD fixed. Thirty days of continuous tracking across six AI platforms, competitor benchmarking, a report of around thirty pages, and a 90-day plan. The most thorough option on this list, priced accordingly.
Pilot Digital — from 3,950 USD one-time, plus an optional tracking tool from 199 USD per month. Four platforms, 20+ prompts, five competitors, and structured data recommendations. Suits companies that want ongoing monitoring bundled from the start.
The engines
Google AI Overviews, ChatGPT and Perplexity.
Three engines because they work differently. ChatGPT reads three or four pages from Bing. Perplexity reads five to ten from its own index. Google uses what already ranks. The same company can be in one and absent from the other two.
The conditions
Logged-out incognito browser, so the answers aren't shaped by anyone's search history. Every run is dated and the country is recorded. Each question is run more than once, because the answers move.
What gets logged
For every question: which companies were named, in what order, and every source the engine cited.
That last part is the useful one. The names in the answer come from a handful of pages, and those pages are usually written by someone else in your category.
What you get
A PDF, around nine pages:
Scorecard — where you appear across the 20 questions and three engines
Competitor map — who takes the answers you're missing from
Source analysis — which pages the engines read for your category
False claims — anything stated about you that isn't true, quoted, with the source
Fix list — what to publish, in the order that moves the most answers
The report tells you what to write and where. It doesn't include the writing. That's the sprint.
Timing
The audit takes about a week from start to delivery.
How the audit works
I run 20 real buyer questions through three AI engines and log every answer and every source they cite. You get a PDF showing where your company is missing, who takes those answers instead, and which pages to fix first
The questions
Not keywords. The questions someone actually types before they buy.
Twenty per category, covering the three shapes a buyer question takes:
the broad one, "best X software 2026"
the price one, "cheapest X for a small team"
the alternatives one, "best alternatives to [the leader in your category]"
The rest cover specific needs — a feature, a company size, an integration, a platform.
Comprehensive audits with tracking
Veza Digital — 4,500 USD fixed. Thirty days of continuous tracking across six AI platforms, competitor benchmarking, a report of around thirty pages, and a 90-day plan. The most thorough option on this list, priced accordingly.
Pilot Digital — from 3,950 USD one-time, plus an optional tracking tool from 199 USD per month. Four platforms, 20+ prompts, five competitors, and structured data recommendations. Suits companies that want ongoing monitoring bundled from the start.
The engines
Google AI Overviews, ChatGPT and Perplexity.
Three engines because they work differently. ChatGPT reads three or four pages from Bing. Perplexity reads five to ten from its own index. Google uses what already ranks. The same company can be in one and absent from the other two.
The conditions
Logged-out incognito browser, so the answers aren't shaped by anyone's search history. Every run is dated and the country is recorded. Each question is run more than once, because the answers move.
What gets logged
For every question: which companies were named, in what order, and every source the engine cited.
That last part is the useful one. The names in the answer come from a handful of pages, and those pages are usually written by someone else in your category.
What you get
A PDF, around nine pages:
Scorecard — where you appear across the 20 questions and three engines
Competitor map — who takes the answers you're missing from
Source analysis — which pages the engines read for your category
False claims — anything stated about you that isn't true, quoted, with the source
Fix list — what to publish, in the order that moves the most answers
The report tells you what to write and where. It doesn't include the writing. That's the sprint.
Timing
The audit takes about a week from start to delivery.
E-mail marek@dioptera.eu
AI visibility audits for software companies.
GEO vs AEO vs SEO
How long until AI answers change
Why your company isn't in ChatGPT
How AI engines pick what to recommend
Do reviews affect AI answers
How the audit works
Comparison pages and AI answers
What an AI visibility audit costs
© 2026 Dioptera
© 2026 Dioptera