Miad Ghazi · Independent SEO consultant

GEO consulting.
Understand your visibility in AI search.

When people ask AI tools about your services or industry, what information do they find? I review the answers, the sources they cite and the pages on your website that could be more useful.

You receive a documented baseline and a practical improvement plan, with a clear distinction between what we observed and what we can reasonably conclude.

Based in Milan. Working remotely with English-speaking teams worldwide.

What GEO means here

Search fundamentals, applied to AI answers

Generative Engine Optimization, or GEO, is a term used for work aimed at improving how a business and its information appear in AI-generated answers. This AI search optimization work includes research, content improvements and technical checks—not a separate set of guaranteed ranking rules.

I start with questions relevant to your buyers and review the answers from the platforms agreed in scope. Google’s AI search features, ChatGPT and Perplexity do not all work the same way, so findings are recorded separately.

It complements SEO

Clear content, accessible pages, internal links and accurate business information matter to both. Traditional SEO is not simply keyword density, and adding schema does not make a page an authoritative source.

Google states that no special AI files or schema are required for AI Overviews or AI Mode. Read Google’s guidance.

Evidence before recommendations

GEO audit: assess your AI search visibility

An AI visibility audit with a defined question set

We agree topics, buyer questions, markets, languages and platforms. The sample includes relevant non-branded questions rather than only prompts designed to mention your company.

A dated visibility baseline

Record tested prompts, platform or mode, language, date, observed mentions and linked sources. Repeat selected checks to understand variation; one answer is not a stable ranking.

A review of cited sources and your pages

Compare the information users receive with your actual offer. Identify inaccurate descriptions, missing explanations, weak evidence and relevant pages that need technical or editorial attention.

A prioritized work plan

Recommendations explain the observation, affected page, proposed action and verification method. Content production, development, outreach and ongoing monitoring are agreed separately.

The output: a documented baseline, source observations, page-level recommendations and a review call. Platforms, sample size, deliverable formats and any paid-tool costs are confirmed before work starts.

From an observation to a page recommendation

Demonstration of the report format; no AI response was tested for this example.
Buyer question“What information should I compare before choosing an industrial maintenance provider?”
Platform and dateA real record identifies platform, mode, language, date and test context. No live observation is asserted here.
Cited sourceRecord the actual URL displayed in a real test. This example attributes no citation to a platform.
Hypothetical gapA service page does not explain the equipment supported or the details needed before an assessment.
Action and checkConfirm facts with the specialist, improve that section and check the published output. Repeat the agreed question set and record variation separately from website traffic.

The question set follows buyer decisions: understanding a problem, comparing options and selecting a provider. Repeated observations use comparable platform, language and test conditions where possible. Specific writing or refresh work can be scoped through content marketing services.

What can change on your site

Make the information easier to find and trust

Content and evidence

Explain the service, product or decision in language your customers use. Add relevant examples, source-backed facts and clear limitations. Where expertise or first-hand experience matters, identify who contributed it and what supports the claim.

Useful headings and self-contained answers help readers navigate the page. They are not a reason to pad every article, force a fixed chunk size or repeat keywords.

Consistent business information

Check that your own pages and relevant public profiles describe the business accurately. Third-party coverage should be earned and relevant; I do not promise Wikipedia placement or paid links as a shortcut to AI citations.

Technical accessibility

Review whether important content and links can be crawled and rendered, whether indexing signals agree, and whether structured data matches the visible page. Broader issues can be handled through a technical SEO audit.

Crawler policies need care

Search crawling, model training and user-triggered retrieval are different uses. We review the relevant platform documentation before proposing changes. A training opt-out must not be treated as a blanket rule about search visibility.

I do not present llms.txt as a universal requirement or a proven route to citations. Any experiment should have a stated purpose and limits.

Measure without overstating

Mentions, citations and customers are different measures

Observed visibility

How often your business is mentioned in the agreed sample, and whether the description is accurate. This is sample-specific, not your share of every AI answer.

Linked citations

Which answers link to your site, and to which pages. Reports distinguish a brand mention from a source link and retain the tested question set.

Traffic and outcomes

Track identifiable referrals and resulting inquiries or sales where measurement permits. Missing referrer data and attribution limits are made explicit.

What a comparison can show

Repeat comparable questions using documented settings and dates. A change in the sample can suggest a direction to investigate; it does not establish that a single edit caused the result.

What it cannot show

Prompt checks are not a census of customer behavior. Google includes AI-feature traffic within its Search Console Web reporting, so that report alone is not a separate AI Overviews performance breakdown.

Working together

A focused project with clear boundaries

  1. 01

    Agree the scope

    Define the audience, questions, markets and platforms. We confirm access, deliverables and the fee.

  2. 02

    Establish the baseline

    Collect dated observations and inspect relevant sources and pages. Identify uncertainty alongside findings.

  3. 03

    Make useful changes

    Prioritize content and technical work. Assign implementation to your team or include support in the agreed scope.

  4. 04

    Review the evidence

    Check deployed changes and repeat the agreed observations. Report visibility separately from business results.

You work directly with me

I’m Miad Ghazi, an SEO consultant based in Milan. My background includes local service websites and multi-region projects. Those SEO experiences inform the investigation; they are not presented as proof of GEO results.

More about my background

Scope-based pricing

GEO reviews, implementation support and ongoing monitoring are custom-quoted. Cost depends on the questions, platforms, markets and work involved. Tool subscriptions or external specialists are discussed before committing to them.

Prices exclude VAT. The initial consultation is free; a full audit and implementation are paid, separately scoped work.

Before we begin

Questions about GEO consulting

Can you guarantee mentions in ChatGPT or Google AI answers?

No. Platforms choose the answers and sources they show, and outputs can vary between runs and users. I can investigate visibility, improve relevant information and document changes—not guarantee citations, rankings or revenue.

How long does GEO take?

Delivery dates depend on the scope and access available. Changes in AI answers have no dependable timetable. Search retrieval and model training are different processes, so I do not tie results to a supposed universal model-update cycle.

Do I need special schema or llms.txt?

Google does not require special schema or AI text files for its AI search features. Appropriate structured data should accurately describe visible content. An llms.txt experiment is not a substitute for accessible pages, useful information or measurement.

Which platforms and tools are included?

We agree the platforms, modes and question sample before starting. Tooling follows the project requirements and budget; a specific paid subscription is not automatically included or required. Results are documented with their collection limits.

Can this cover several countries or languages?

Yes, if those markets and language-review requirements are included in scope. Questions, available features and cited sources may differ by locale. International targeting and localization may also need a separate international SEO workstream.

Should we start with GEO or a technical SEO audit?

If important pages cannot be crawled, indexed or used properly, address those fundamentals first. If the site is accessible and you need to understand how the business appears in AI answers, a focused GEO baseline can help define the next step.

Planning several markets? See international SEO consulting.

Start with your question

What would you like to find out?

Send your website address, the audience you serve and the questions you want to investigate. We’ll discuss whether a focused review is useful for your business.

Discuss your AI search visibility

Free initial consultation. Scope and fee agreed before work begins.