This is a template, not the briefing
Every session I run is tailored to the client: your category, your markets, your numbers and the decisions on the table. In the live briefing the examples and data are yours, not these generic ones.
Use this deck to see the approach. The live briefing, built around your business, is a different thing.
WHY NOW — THE SHIFT IS REAL
The search surface changed. Most content hasn't yet
Search is no longer just ranking on Google or showing up on Amazon. It's being selected, summarised and recommended by AI systems — and content built for human reading and classic SEO is often invisible to them.
monthly users see Google AI Overviews in search results
people use ChatGPT every week — many of them instead of a search engine
fewer clicks on search results when an AI summary appears (8% vs 15% of visits)
SOURCE Pew Research Center, 2025
customers used Amazon's Rufus shopping assistant
WHAT IT LOOKS LIKE ON ONE SITE — illustrative: clicks cut in half after AI Overviews · AI answers trigger on 5,300 of 7,600 keywords · yet the site is cited in only 330 of them.
How an AI answer gets built
Simplified, but this is the logic behind Google AI Mode, ChatGPT search and similar tools. AI doesn't rank pages — it assembles one answer from passages.
“My hands crack every winter and I wash them a lot at work. What cream actually helps?”
Long, personal, situational
Why skin cracks · key ingredients · creams for frequent washing · how fast it works · when to see a doctor
Often 5–20 searches run in the background
For each sub-question, the model takes the smallest piece of text that answers it — from experts, retailers, brands and forums.
A page with no clear answer is skipped
The passages are combined into one answer that names a few products with reasons, and links 3–6 sources.
Only cited sources get the click
WHAT IT MEANS: passages compete, not pages. A clear answer at the start of a section is easier to use than a better one buried further down.
WHAT VISIBILITY MEANS IN AI
Mentioned, cited, recommended
A brand can appear in an AI answer in three different ways. The most valuable is a recommendation — and it always comes with reasons.
The model gives a verdict — “a good first choice is…” — and explains why it fits the person's situation.
The goal. It only happens when the reasons are easy to find.
The brand appears in a list or comparison, but not as the first choice.
Visible, but the decision goes to someone else.
A brand page is linked as evidence. Being cited and being named are separate: a site can be quoted without the brand being recommended.
Builds trust, and the click comes from here.
KEY POINT: AI recommends a product when it can explain why. If your content doesn't state the reasons clearly, the model takes them from someone else — or picks a product whose reasons are easier to find.
These three outcomes are what AI visibility tools measure — the basis of before / after reporting.
WHERE AI ANSWERS COME FROM
Three sources, three levels of control
AI answers draw on several places at once. In sensitive categories, the model checks whether others confirm what the brand says before it recommends anything.
Product pages, guides, buying advice, FAQs.
You write and publish it. The place AI goes to confirm product facts and reasons.
Retailer & marketplace pages
Online stores, chains and marketplaces: titles, descriptions, Q&A, reviews.
You supply the content; retailers decide what is shown. Key for “what can I buy” questions.
Experts, associations, review sites, media, forums.
Others write it. Often carries the most weight — brand content wins when it agrees with them.
WHERE PEOPLE SEE THE ANSWER
ChatGPT · Gemini · Perplexity · Copilot
KEY POINT: no single source wins on its own. Brand pages, retailer listings and independent voices need to tell the same, clear story.
Same facts, new structure
The work changes how content is structured — never the information it contains.
Existing content goes in; AI-ready content comes out, with identical facts.
Your current pages, product cards & articles — facts, claims and data exactly as they are today.
Answer-first order, question-led headings, FAQ blocks & schema. AI-assisted, human-checked; the facts are not rewritten.
Optimised, AI-ready content
Extractable pages — same facts, same message, now readable and citable by AI systems.
Structure · heading logic · answer placement · FAQ blocks · schema.org · extractability
Facts · claims · data · brand message · legal and regulatory accuracy · meaning
Same facts in, same facts out — so brand and legal review becomes faster and lighter: reviewers check structure and wording, not new claims.
as optimised, well-written content for your team to approve.
Start with the 15 pages that matter most
Not every page is worth the same. Optimise a focused priority set first — where AI answers already appear and the brand has the most to win — then scale with results in hand.
HOW TO PICK THE PRIORITY SET
AI answers already appear
Queries where AI Overviews and chatbots answer today.
Topics where competitors or retailers get quoted, not you.
Hero products, key categories, pages closest to purchase.
One page per template, so each becomes a reusable blueprint.
15 priority pages, rebuilt to the content guidelines
AI visibility & citations, before vs after
Roll the proven blueprints out to the rest of the site
I set the standard, your teams own the content
Frameworks and tools are built once. Each market or brand then gets its own audit and content work, so local voice, products and regulatory nuance stay local.
Built once, for all markets
• AI search fundamentals overview
• Content guidelines for GEO
• Technical GEO guidelines (schema, HTML, crawlers)
• Prompt selection guidelines
• Training / workshop for local teams
• Schema templates
TOMEK SIKORA · PER MARKET
• Competitive AI visibility benchmark
• Prompts review
• Technical audit & fixes in the CMS
• Structured data on key URLs
• Content audit & optimisation of the priority pages
• Before / after report
Owns decisions and approval
• Priority areas & hero products
• Access to tools and the CMS
• Brand & legal review and approval
• Go-live sign-off
WHY SUPPORT, NOT LEAD: no upfront master content before the first market goes live, a natural fit for markets with different products and rules, and one shared toolset keeps quality and measurement consistent.
Same questions, before and after
Every market is measured the same way: a fixed set of questions people ask AI, checked before the changes and again after a month of data.
Unbranded questions people really ask, each mapped to a priority page. Kept unchanged for the whole test.
Where the brand is recommended, mentioned or cited today — per AI engine, vs competitors.
The 15 priority pages, rebuilt to the guidelines and approved by brand & legal.
After 1 month of data: the same questions, the same engines, the same competitors.
AI VISIBILITY AI visibility tracker
How often and how AI models present the brand vs competitors — recommended, mentioned or cited, per model and per market.
SEARCH & TRAFFIC Search Console & GA4
Impressions, clicks, CTR and visits on the optimised pages, before vs after.
RESULT: a before / after report per market, comparable across markets — the basis for deciding where to scale.
Who this session is for
Leadership, marketing directors and the people who decide budgets and priorities. Thirty minutes, no jargon: what changed, why it matters for revenue, and a priority-first plan that proves results on a few pages before anything is scaled.
It is one module in my SEO & AI Search training. Teams that will do the work can continue with AI search: the basics.
