DEV Community: Łukasz B.

How AI search picks fragments: query fan-out and RAG explained

Łukasz B. — Wed, 07 Oct 2026 22:21:01 +0000

Hi again 👋 In my last post, How to become an SEO (AI) specialist in the coming 2027 #2, I wrote that without understanding the mechanics of AI search there is no "AI SEO", only guessing. A lot of you asked me to go deeper. So today we open the hood 🔧

In this post I will explain, step by step, what happens between the moment a user types a question into AI Mode, ChatGPT or Perplexity and the moment an answer with citations appears on the screen. And, most importantly, where in that process your content wins or loses.

Why an SEO needs to understand RAG
The big picture: from question to answer
Step 1: query understanding and fan-out
Step 2: retrieval
Step 3: reranking and fragment selection
Step 4: generation, grounding and citations
Not every AI searches the same way
What this means for SEO in practice
Myths I keep seeing on LinkedIn

🤔 Why an SEO needs to understand RAG

RAG stands for Retrieval-Augmented Generation. In plain words: the model does not answer only from memory. First it searches, then it reads what it found, and only then writes an answer based on it.

Almost every AI search product works this way: Google AI Overviews and AI Mode, ChatGPT Search, Perplexity, Copilot, Gemini. The details differ, but the skeleton is the same.

Why does it matter for us? Because classic SEO optimized for one moment: ranking a URL for a keyword. RAG has four moments where your content can drop out:

the system never asks a question your page answers (fan-out),
your page is not retrieved for that question (retrieval),
your page is retrieved, but another fragment scores higher (reranking),
your fragment is read, but the model does not cite it (generation).

Ranking #1 helps with only some of these. That is why you can rank first and still be invisible in the AI answer 🤯

🗺️ The big picture: from question to answer

Before we go into details, here is the whole pipeline in one place. Every AI search system has its own version, but they all follow roughly these steps:

Query understanding. The model reads the question, the conversation history and the context (location, language, sometimes the user's past chats).
Fan-out. One question becomes several or many sub-queries.
Retrieval. Each sub-query goes to the index (web, Knowledge Graph, shopping, news, maps) and returns candidates.
Chunking and reranking. Candidate pages are cut into fragments, and a stronger model picks the best ones.
Generation. The LLM gets the question plus the selected fragments and writes the answer.
Citation. The system attaches sources to the claims in the answer.

And here is what decides whether you win at each stage:

Stage	What it looks for	How your content wins
Fan-out	the questions behind the question	you cover the sub-topics users really care about
Retrieval	documents that match a sub-query	the page is indexed, accessible and textually relevant
Reranking	the best fragment for a sub-query	a paragraph that answers precisely, on its own
Generation	facts the model can use	clear, specific, verifiable statements
Citation	a source that supports a claim	your fragment is the one that actually contains the fact

Now let's go through it step by step 👇

🌿 Step 1: query understanding and fan-out

Google states it directly in its Search Central documentation: both AI Overviews and AI Mode may use a "query fan-out" technique, issuing multiple related searches across subtopics and data sources to build a response.

In practice it looks like this. A user asks:

"Which running shoes are best for a beginner with flat feet who runs on asphalt?"

The model does not search for that sentence. It breaks it into sub-queries, for example:

running shoes for flat feet (stability vs neutral)
best beginner running shoes 2026
cushioning for asphalt running
reviews of specific models
how to choose running shoe size

Each of these goes to the index separately. The answer is then assembled from fragments that answered different sub-queries 🧩

A few things worth knowing:

Not every query fans out. A simple fact ("capital of Spain") does not need it. Complex, comparative and "it depends" questions fan out the most.
The sub-queries are hidden. You will not see them in Search Console, and no third-party tool has access to Google's internal ones. Tools that "show fan-out queries" are simulating them with their own LLM. Useful as inspiration, not as data.
Fan-out is not query rewriting. The original question still matters. The system searches it and the sub-queries.
Context changes the fan-out. Location, language, conversation history and personalization mean two users with the same question can trigger different sub-queries.

💡 The SEO consequence: a page can be cited for a question it does not rank for, because it was the best answer to one of the sub-queries. And the opposite: your #1 page for the main keyword may not answer any sub-query precisely enough.

🔎 Step 2: retrieval

Retrieval is the moment the system goes to its index and brings back candidates for each sub-query. Two important facts first:

AI does not search the live web from scratch. It searches an index: Google's own, Bing's, or the AI company's own crawl. If you are not in that index, you do not exist. For Google's AI features the rule is simple: the page must be indexed and eligible to show with a snippet. Nothing more, nothing less.
What the bot could not read, the index does not have. Content loaded by JavaScript that a given bot does not render, content behind a login or blocked in robots.txt never reaches retrieval.

Chunking: pages become fragments

LLMs do not work on whole pages. Pages are split into chunks: paragraphs, sections or fixed-length pieces of text. Each chunk is scored separately. Your 4,000-word guide is, for the retrieval system, maybe 30 separate candidates competing with each other and with the rest of the web.

Lexical vs semantic search

There are two main ways to find matching chunks:

Method	How it works	Strength	Weakness
Lexical (e.g. BM25)	matches the words in the query to the words in the text	precise for names, numbers, product codes	misses synonyms and paraphrases
Semantic (embeddings)	turns query and text into vectors and compares meaning	finds an answer even with different wording	can miss exact terms and rare names
Hybrid	combines both scores	the standard in modern systems	—

An embedding is a list of numbers that represents the meaning of a text. Texts with similar meaning have similar vectors, so "cheap flights to Rome" and "budget airfare Italy capital" land close to each other, even though they share almost no words.

💡 The SEO consequence: keywords did not die, they just stopped being enough. Exact terms (brand names, model numbers, prices, places) still help lexical matching. Clear, topical paragraphs help semantic matching. A paragraph that mixes three topics produces a "blurry" vector that matches nothing well.

🏆 Step 3: reranking and fragment selection

Retrieval is fast but rough. It may return hundreds of candidate chunks per sub-query. The model generating the answer cannot read them all, because its context window is limited and every extra token costs money and time ⏱️

So there is a second, more precise stage: reranking. A stronger (and slower) model looks at the sub-query and each candidate fragment together and scores how well that specific fragment answers that specific question. Only the top few go further.

What typically helps a fragment survive reranking:

It answers the question directly, preferably in the first sentence.
It is self-contained. It makes sense without the paragraph above it. "As mentioned earlier, it costs 20% less" is a weak chunk, because what costs less?
It is specific. Numbers, names, dates, conditions. "It depends" without explaining on what loses.
It comes from a source the system trusts. Many systems mix relevance with quality and authority signals at this stage. Google has its whole classic ranking toolkit available here.
It is fresh, when freshness matters. For prices, versions and news, a newer fragment often wins.

Here is the difference in practice:

❌ Weak chunk	✅ Strong chunk
"There are many factors to consider and every runner is different, so it's best to try a few options."	"Runners with flat feet usually do better in stability shoes, which limit inward rolling of the foot (overpronation). For a beginner on asphalt, choose a stability shoe with medium-to-high cushioning."

💡 The SEO consequence: the unit of competition is now a paragraph, not a page. Your page can be great overall and still lose every individual "duel" with sharper fragments from other sites.

✍️ Step 4: generation, grounding and citations

Now the LLM gets a prompt that roughly says: here is the user's question, here are the selected fragments, write an answer based on them. This is called grounding: the answer is anchored in retrieved sources instead of only in the model's memory.

A few mechanics that matter for us:

1. The model synthesizes, it does not copy. It combines facts from several fragments into its own sentences. That is why your exact wording rarely appears in the answer, but your facts can.

2. Citations are attached to claims. The system links sentences in the answer to the fragments that support them. A fragment that contains a clear, checkable fact ("the battery lasts 30 hours") is easy to cite. A fragment full of opinions and fluff gives the model nothing to attach a citation to.

3. Read is not the same as cited. The model may read your fragment, use it as background and still cite someone else. Interfaces show a limited number of sources, so there is competition even at this last step.

4. Consensus wins. When many independent sources say the same thing, the model treats it as safe. When your fragment contradicts the majority, it needs strong support (data, a source, first-hand evidence) to be used.

5. Memory still leaks in. Grounding reduces hallucinations, but does not remove them. The model's training knowledge can still shape the answer, which is how outdated prices or old offers about your brand end up in AI answers ⚠️

💡 The SEO consequence: write fragments that contain citable facts. Original data, specific numbers, clear definitions and first-hand experience are exactly what a model needs to justify a sentence, and what it cannot get from everyone else.

🔀 Not every AI searches the same way

The pipeline is similar, but each system has its own index, its own bots and its own taste in sources. That is why your brand can look great in Perplexity and be missing in AI Overviews.

System	Where it retrieves from	What it means for you
Google AI Overviews / AI Mode	Google's index (crawled by Googlebot), Knowledge Graph, Shopping, Maps	classic Google SEO is the entry ticket; no extra files or markup needed
ChatGPT Search	OpenAI's own crawler (OAI-SearchBot) plus partner search providers	blocking OAI-SearchBot removes you from ChatGPT search results
Perplexity	its own index (PerplexityBot) plus live fetching	usually shows many sources, so niche pages have a real chance
Copilot	Bing's index	Bing Webmaster Tools and IndexNow are worth your time again

Memory vs live search

There is one more distinction many people miss. An assistant can answer in two modes:

From training data (memory). No search happens, no citations appear. The answer reflects what the internet said about you when the model was trained, sometimes a year or more ago.
With retrieval (search). The pipeline from this post runs, and citations appear.

The model often decides on its own whether to search. Questions about current prices, news, comparisons and "best X in 2026" usually trigger search. General questions ("what is CRM software?") are often answered from memory.

💡 The SEO consequence: you need two strategies. Retrieval you influence with technical SEO and strong fragments, today. Memory you influence slowly, through a consistent picture of your brand across many sources that end up in future training data.

✅ What this means for SEO in practice

Theory is nice, but here is the checklist I actually use. One block per pipeline stage 👇

Fan-out: cover the questions behind the question

List the sub-questions a real user has around your topic: comparisons, costs, risks, "for whom", "how to choose", alternatives.
Check People Also Ask, Reddit threads, sales and support questions. They are the closest thing to real fan-out data you have.
Cover the sub-topics in one strong, well-structured page where it makes sense, not in 50 thin pages. Google warns that mass-producing pages for every query variation can fall under its scaled content abuse policy.

Retrieval: make sure you are in the index

Check robots.txt and your CDN/WAF for Googlebot, OAI-SearchBot, PerplexityBot and Bingbot. Verify in the server logs, not only in the file 🔍
Make key content (descriptions, prices, specs, FAQs) available in the raw HTML, not only after JavaScript runs.
Use the words people use: product names, model numbers, locations. Lexical matching still counts.

Reranking: write fragments that win duels

Start each section with a direct answer, then expand.
Make each paragraph self-contained: one topic, no "as mentioned above".
Use descriptive headings that sound like the sub-question they answer.
Replace "it depends" with the conditions it depends on.

Generation: give the model citable facts

Add numbers, dates, definitions and conditions that can support a sentence.
Publish original data, tests and first-hand experience. That is the one thing other sites cannot copy.
Keep facts about your brand (prices, offer, locations) consistent and current everywhere: your site, Google Business Profile, comparison sites, reviews.

Measurement

Track your brand's citations for key questions in AI Overviews, ChatGPT and Perplexity. Treat the results as a trend, not a rank, because answers are non-deterministic.
Segment AI referral traffic in GA4 (chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com).

🙅 Myths I keep seeing on LinkedIn

Myth	Reality
"You need llms.txt or special AI markup to show up in AI Overviews."	Google says no new machine-readable files, AI text files or special schema are needed. Indexing and eligibility for a snippet are the requirement.
"Create a separate page for every fan-out query."	Fan-out queries are hidden, and mass-producing pages for query variations risks scaled content abuse. Cover sub-topics well instead.
"Rank #1 and you'll be cited."	Ranking helps retrieval, but reranking and generation work on fragments. #1 pages are regularly not cited.
"This tool shows Google's fan-out queries."	It shows queries generated by the tool's own LLM. Useful for ideas, not a data source.
"Short chunks of 40–60 words always win."	There is no universal ideal length. A fragment wins by being complete and precise, not by a word count.
"AI answers from memory, so SEO doesn't matter."	Answers with citations come from retrieval, and retrieval runs on indexes built by crawlers. Technical SEO is the entry ticket.

🧾 Summary

AI search is not magic. It is a pipeline: fan-out → retrieval → reranking → generation → citation. At every stage your content can win or drop out, and each stage rewards something slightly different:

fan-out rewards covering the real questions behind the query,
retrieval rewards being indexed, accessible and textually relevant,
reranking rewards precise, self-contained paragraphs,
generation rewards specific, citable facts and a consistent brand picture.

The good news? None of this replaces SEO fundamentals. It sits on top of them. The specialists who understand the pipeline stop guessing and start testing 🧪

In my next post (#3) I will show step by step how to measure a brand's visibility in AI answers and how to turn it into a client report.

About the author ✍️

Łukasz, SEO AI Specialist at Afterweb. He has worked in SEO for years and today focuses on brand visibility in Google and in AI answers: from technical SEO, through data analysis and automation, to content and brand-mention strategy. He shares what he knows in the "How to become an SEO specialist" series, because he believes the best way to learn is practice and testing on real data.

Want to talk about SEO and your brand's visibility in AI? Visit afterweb.pl 🚀

Sources

AI features and your website, Google Search Central
Query Fan-Out Technique in AI Mode: New Details From Google, Search Engine Journal
Query Fan-Out Rewards One Page That Answers Many Questions, Xponent21

🔥 How to become an SEO (AI) specialist in the coming 2027 #2

Łukasz B. — Thu, 01 Oct 2026 21:09:05 +0000

Hi again 👋 Four years ago I wrote my first post here: How to become an SEO specialist in the coming 2022 #1. Today I'm back with part two, because a lot has changed since then. Honestly? Half of the skills from that list belong in a museum 🏛️, and the other half is barely enough to get started.

In this post I will discuss what a person needs in 2027 to start (and grow) as an SEO (AI) Specialist.

1. What happened since my 2022 post
2. Who is an SEO (AI) Specialist in 2027
3. What died and what was born (2022 vs 2027)
4. How AI search actually works
5. Where you can work as an SEO (AI) Specialist
6. Skills at the Junior level
7. Skills at the Senior level
8. The SEO (AI) specialist's toolkit
9. Everyday tasks in 2027
10. Roadmap: your first 6 months
11. Will AI replace SEO specialists?

🕰️ What happened since my 2022 post

In 2022 we worked with Universal Analytics, Google My Business and FID. We wrote "SEO texts" around keywords and reported top 10 rankings. The word "AI" showed up mostly in tool vendors' sales decks.

In 2026 users ask their questions in AI Overviews, AI Mode, ChatGPT, Perplexity or Gemini, and often get an answer without a single click. Search is no longer a list of ten links. It is a system that assembles an answer from fragments of many sources 🧩

That changes the question itself. It is no longer "how do I get into the top 3", but "how do I make my brand the source AI picks, cites and recommends". This post is for people who want to answer that question professionally in 2027.

🤖 Who is an SEO (AI) Specialist in 2027

An SEO (AI) specialist is responsible for whether a brand is visible, cited and recommended wherever people look for information. That means Google (classic results, AI Overviews, AI Mode), AI assistants (ChatGPT, Perplexity, Gemini, Claude, Copilot) and platforms such as YouTube, TikTok, Reddit or Amazon.

In 2022 a "positioner" optimized a website for an algorithm. In 2027 an SEO (AI) specialist optimizes the brand's entire online presence for systems that read, summarize and recommend. The website is important, but it is only one of the sources AI uses to build its picture of a company.

New roles have appeared that did not exist in 2022 🆕

Role	Responsible for
AI Search / GEO Specialist	brand visibility and citations in AI answers
SEO Automation Specialist	scripts, data pipelines and AI agents for SEO work
SEO Data Analyst	GSC, GA4, BigQuery, server logs, forecasting models
Technical SEO / JS SEO	rendering, crawl budget, content accessibility for AI bots
Content & Entity Strategist	topics, entities, E-E-A-T, original data and research
Digital PR / Brand Mentions	brand mentions in the sources AI cites

The classic career path (Junior, Mid, Senior, Head of SEO) still exists 💰 The difference is that from Mid level up, you are now expected to do part of the work through automation and AI, not by hand.

⚰️ What died and what was born (2022 vs 2027)

The fastest way to see how much has changed is to put my 2022 list next to today's reality.

Area	2022	2027
Analytics	Universal Analytics, GAIQ certificate	GA4, Looker Studio, BigQuery export
Local SEO	Google My Business	Google Business Profile + reviews and mentions across the web
Core Web Vitals	FID	INP (since March 2024)
Content	"SEO texts" written around keywords	content with experience, original data and E-E-A-T
AI in content	a curiosity	a work tool; mass content with no value = scaled content abuse
Links	link building for rankings	links + brand mentions in sources cited by AI
Measuring success	rankings and organic traffic	rankings, traffic, AI citations, share of voice, conversions
Frontend	AngularJS, Backbone	JS SEO: SSR vs CSR, rendering for AI bots
Bots	Googlebot, Bingbot	+ GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot and others
Search	10 blue links	AI Overviews, AI Mode, AI chat answers, zero-click

The takeaway is simple: the technical SEO fundamentals are still here, but what search rewards and how we measure results have changed. If you do SEO in 2027 the way you did in 2022, you are doing it for a shrinking share of traffic 📉

🧠 How AI search actually works (and why it is the key to SEO)

Without understanding the mechanics there is no "AI SEO", only guessing. Here are five things a specialist has to understand at a technical level.

1. Query fan-out. AI systems (e.g. Google AI Mode) break a single user question into many sub-queries and search for answers to each of them separately. A page can therefore be cited even though it does not rank for the main keyword, because it answers one of the sub-queries. It can also rank #1 and not be cited at all 🤯

2. Fragments get cited, not whole pages. Models retrieve and score pieces of text (passages, chunks). The winner is a paragraph that is a complete, precise answer on its own, not a page that is "generally about the topic".

3. Most AI crawlers do not render JavaScript. Googlebot renders JS, but many AI assistants' bots fetch only the raw HTML. If your content, prices or product descriptions load client-side (CSR), they simply do not exist for those bots. This is one of the most frequently missed issues in audits ⚠️

4. Not every AI bot does the same job. You have to tell training bots apart from search bots and from bots acting on a user's request:

Bot	What it is for	What happens if you block it
GPTBot	training OpenAI models	content stays out of training; ChatGPT Search visibility unchanged
OAI-SearchBot	ChatGPT search index	the site disappears from ChatGPT Search results
ChatGPT-User	fetching a page at a user's request	ChatGPT will not open the page when a user asks for it
Google-Extended	using content to train Gemini	does not turn off AI Overviews or AI Mode (those use Googlebot)
PerplexityBot	Perplexity index	no citations in Perplexity

One wrong line in robots.txt can cut a brand out of ChatGPT without the client ever knowing. An SEO (AI) specialist has to be able to verify this in the server logs, not just in the file 🔍

5. AI trusts what "the internet says" about a brand. Models do not rely on your website alone. They take into account reviews, comparison sites, discussions on Reddit and forums, YouTube videos and industry articles. A consistent picture of the brand (the entity) across many independent sources is now as important as your link profile.

🏢 Where you can work as an SEO (AI) Specialist

The three paths from 2022 still exist: agency, in-house and freelance. What each of them expects from you has changed, though.

Path	What you gain	What changed since 2022
Agency	many projects, fast learning from case studies	smaller teams; one specialist handles more clients thanks to automation, and clients ask about their visibility in ChatGPT
In-house	deep knowledge of the brand, influence on product and data	SEO works with PR, content, product and analytics; what counts is impact on revenue, not just traffic
Freelance	freedom, choice of clients, your own pace	with AI and automation one person does the work of a small team; specialists with a narrow niche win

A new, fourth option is product companies and SaaS built around SEO and AI 🚀: tools for monitoring LLM visibility, content platforms and rank trackers. They are looking for people who understand SEO, data and product at the same time.

🌱 Skills at the Junior level

In 2022 I gave Juniors a list of almost 30 items. This time there are priorities: first what you cannot work without, then the rest.

Must-have (the foundation):

How a search engine works: crawling, rendering, indexing, ranking. Plus the same for AI search: retrieval, query fan-out, citations.
Technical basics: HTTP status codes, redirects, robots.txt, meta robots, canonical, XML sitemaps, hreflang, structured data (schema.org).
Google Search Console at an everyday level: indexing reports, performance, URL inspection.
GA4: events, conversions, traffic sources, including traffic from chatgpt.com, perplexity.ai and gemini.google.com.
A crawler (Screaming Frog or Sitebulb): a full audit of a small site.
Intent analysis: not just "which keyword", but "which question and which answer format".
HTML and JS basics: so you understand what the bot sees versus what the user sees.

New in 2027 (you can't start without these anymore):

Working with LLMs: writing prompts for analysis, briefs and keyword classification, but above all verifying what the model returns. A Junior who copies an AI answer without checking it is a risk to the team.
Checking brand visibility in AI: how the brand shows up in ChatGPT, Perplexity and AI Overviews for key questions, and who gets cited instead.
E-E-A-T in practice: authors, sources, first-hand experience in the content.
Looker Studio: a simple dashboard built on GSC and GA4.

Nice to have: Python or Google Apps Script basics, Google Business Profile, WordPress or Shopify at a power-user level.

🦾 Skills at the Senior level

A Senior in 2027 is not defined by how many tools they know. They are defined by whether they can translate SEO into business decisions and automate what is repetitive.

Strategy and business

Visibility strategy: classic SEO, AI search, brand mentions and platforms (YouTube, Reddit, marketplaces).
Traffic and revenue forecasting, plus a business case for the board.
Reporting in the zero-click era: share of voice, AI citations, branded traffic and conversions instead of organic traffic alone.

Data and automation

Python: APIs (GSC, GA4, SEO tools, LLMs), pandas, scraping, report automation.
SQL and BigQuery: working on the raw GSC and GA4 export.
Embeddings: keyword clustering, cannibalization detection, internal linking based on semantic similarity.
AI agents and pipelines: automated audits, change monitoring, brief generation with quality control.

Expert-level technical SEO

Server log analysis: Googlebot and AI bots, i.e. who fetches what and how often.
JS SEO: SSR, SSG, hydration, content availability in the raw HTML.
Crawl budget and information architecture for large sites and e-commerce.
AI bot policy: deliberate decisions about which bots to let in, which to block, and why.

Content, entities and brand

Passage-level optimization, topic and entity architecture.
Digital PR focused on the sources AI actually cites in a given industry.
Original research and data as an advantage nobody can copy with a prompt.

People

Leading a team, running training sessions, working with developers, product and PR.
The ability to say "I don't know, I'll test it" instead of repeating LinkedIn myths 😉

🧰 The SEO (AI) specialist's toolkit

Tools change, categories stay. Know one tool from each category and understand what data it actually returns.

Category	Examples	What for
Google data	Google Search Console, GA4, BigQuery	the source of truth on traffic and queries
Crawlers and audits	Screaming Frog, Sitebulb	technical health, rendering, structure
SEO suites	Ahrefs, Semrush, Senuto (Polish market)	keywords, links, competitors
AI visibility	Ahrefs Brand Radar, Semrush AI Toolkit, Profound	brand citations and mentions in AI answers
Server logs	Screaming Frog Log File Analyser, custom scripts	Googlebot and AI bot activity
LLMs for work	ChatGPT, Claude, Gemini	data analysis, scripts, briefs, classification
Automation	Python, n8n, Make, Apps Script	reports, monitoring, data pipelines
Reporting	Looker Studio	dashboards for clients and the board

The most important rule: an AI monitoring tool shows a sample of prompts, not the whole truth. LLM answers are non-deterministic and depend on the user, location and context. Treat this data as a trend, not as a position in a rank tracker 📊

📋 Everyday tasks in 2027

When I compare my 2022 task list with today's, one trend stands out: less manual work, more analysis, automation and brand building.

Task in 2022	What it looks like in 2027
manual keyword research	clustering thousands of keywords with embeddings and mapping intents and fan-out questions
SEO audit in a spreadsheet	automated crawling and monitoring + an audit of content accessibility for AI bots
monthly ranking report	a dashboard: traffic, conversions, share of voice and AI citations
writing SEO texts	data-driven briefs, quality control of AI-assisted content, adding first-hand experience
link building	digital PR and mentions in sources cited by AI
working with developers	collaborating on SSR, structured data and bot policy
client training	training on AI search and on how to read the new metrics

New tasks that did not exist at all in 2022:

monitoring AI answers to key questions about the brand and its competitors;
analyzing logs for AI bots and making robots.txt decisions;
building and maintaining automations and AI agents in the team;
reacting when AI states false information about the brand (e.g. old prices or an outdated offer).

🗺️ Roadmap: your first 6 months

Theory from courses is not enough. This plan is built on one principle: learn on your own website, one you are allowed to break.

Month 1: your own sandbox. Launch a site (WordPress or a static site generator), connect GSC and GA4, write 5 articles based on your own experience. Learn to read the indexing reports.
Month 2: technical SEO in practice. Crawl the site with Screaming Frog. Deliberately break a canonical, a redirect and robots.txt, then watch what happens in GSC. Add structured data 🔧
Month 3: AI search. Check how your niche looks in AI Overviews, ChatGPT and Perplexity: who gets cited and why. Rewrite one article for passages (the answer at the start of each section) and compare the result.
Month 4: data and logs. Pull your server logs and find Googlebot and AI bots in them. Build your first Looker Studio dashboard.
Month 5: your first code. Write a Python script that pulls data from the GSC API and flags the pages with the biggest drop in clicks. Use an LLM to help, but understand every line 🐍
Month 6: case study. Describe what you did, what the data looked like before and after, what worked and what didn't. Publish it on a blog or LinkedIn. That is your CV.

After six months you have something most course graduates don't: your own project, your own data and proof that you can think like a specialist 💪

🤔 Will AI replace SEO specialists?

Short answer: AI won't replace SEO specialists. It will replace the ones who still do SEO as if it were 2022.

AI already writes meta descriptions, groups keywords, generates briefs and runs preliminary audits faster than a human. If your job consists mainly of those tasks, you have a real problem.

But AI will not take over the things that decide the outcome:

Judgment. A model doesn't know which of 200 audit recommendations matters for the client's business.
Accountability. Someone has to make the call on a migration, a bot block or an architecture change, and own it.
Experience and data that aren't on the web. What you learn from your own tests is exactly what AI looks for to cite.
People. Convincing the board, getting developers on board and building relationships with the media.

The 2027 paradox: the more content AI generates, the more valuable becomes whatever AI cannot generate. An SEO (AI) specialist uses AI as leverage, not as a replacement for thinking.

✅ Summary

Becoming an SEO (AI) specialist in 2027 is harder than it was in 2022, but also more exciting than ever. It all comes down to three things:

Technical SEO fundamentals are still mandatory, and they are also the foundation of AI visibility.
Understanding AI search (fan-out, passages, bots, brand mentions) is what separates specialists from "positioners" today.
Data and automation decide whether you are a Senior or someone doing tasks AI will do faster.

The patience I wrote about in 2022 is still needed. Add curiosity and a habit of testing, because in 2027 the rules of the game change every quarter, not every few years.

Start today ❗️❗️ 🤓

In my next post (#3) I will show step by step how to measure a brand's visibility in AI answers and how to turn it into a client report.

About the author ✍️

Łukasz, SEO AI Specialist at Afterweb. He has worked in SEO for years and today focuses on brand visibility in Google and in AI answers: from technical SEO, through data analysis and automation, to content and brand-mention strategy. He shares what he knows in the "How to become an SEO specialist" series, because he believes the best way to learn is practice and testing on real data.

Want to talk about SEO and your brand's visibility in AI? Visit afterweb.pl 🚀

🔥How to become an SEO specialist in the coming 2022 #1

Łukasz B. — Sun, 24 Oct 2021 20:57:02 +0000

Hi, My name is Łukasz and it's my first post here 😇. I would like to share my knowledge of working as an SEO Specialist. In this post I will discuss the topic that I see such a person should have in 2022 to start work. The development of technology is progressing very quickly, which is why the requirements set for SEO specialists too.

1. Who is an SEO Specialist
2. Where can we work as an SEO Specialist
3. What skills should an SEO Specialist have
4. Everyday tasks in the SEO industry

Who is an SEO Specialist

An SEO specialist is a person who works with the optimization and marketing of a website in order to achieve higher search engine rankings (commonly referred to as a website positioner).
The word SEO specialist is not related to just one kind of job.

The SEO industry is extremely broad 📏- apart from several career levels of an SEO Specialist, we also distinguish positions that include people dealing with content marketing, copywriting, linkbuilding or customer contact. So we can list jobs: SEO Analyst, SEO Specialist, SEO Consultant, SEO Technician, SEO Manager, SEO Trainer, SEO Director, and the position of SEO Specialist can be divided into career paths - Junior SEO Specialist, SEO Specialist, Senior SEO Specialist. Head of SEO, SEO Team Leader, Full Stack SEO, Full Stack SEO Expert or Full Stack SEO Developer 💰

Where we can work as an SEO Specialist.
Working as an SEO Specialist, we can work in an internet agency, in-house or as a freelancer.

The great advantage of working in an SEO agency is the variety of projects and the exchange of experiences based on various case-studies.

Working in-house, we usually constitute a one-person department whose task is to support the management board and the marketing department in the effective promotion of the company's offer. On the other hand, however, when working for one brand, we have much more time to get to know it, maximize its promotion and the freedom to experiment with new solutions.

When working as a freelancer, we certainly have the greatest freedom in our actions, we choose our clients and organize our working day. Soft skills, good communication and, above all, very good organization of work and self-discipline are very useful here. This type of work also does not require daily contact with clients or colleagues.

What skills should an SEO specialist have
In my opinion, a good SEO specialist should have some of the features:

Technical SEO skills
ON-site SEO skills
OFF-site SEO skills
Data analysis skills
Project management skills
Web development skills
Communication skills

In fact, the best SEO specialists are people who are proficient in many different areas of digital marketing. This includes basic SEO skills, but also digital marketing skills such as copywriting and website design. At the beginning of your career as an SEO Specialist, you should definitely start with gaining knowledge about the secrets of SEO, positioning and website optimization. On the Internet you will find hundreds of interesting courses, both for beginners and more advanced people.

While theoretical knowledge is not the most important for an SEO specialist, it does provide important preparation at the beginning of your career.

On the theoretical level, an SEO specialist should know the basics of computer and network operation, know the basics of creating a website and be fluent in tools such as Excel or Docs. In addition, it is advisable for an SEO specialist to have a basic understanding of web technologies. It is difficult to list all the skills that a good specialist should have, but I will try to specify the appropriate elements according to the experience. You can read more about SEO on the website: https://afterweb.pl/pozycjonowanie-stron/

Skills at the Junior SEO Specialist level:

Computer skills and network
What is a website,Pagerank, proxy server, structural data
What are the URL parameters in the Google search engine
What is Hosting, Domain
FTP client (what is FTP, Scure FTP)
HTTP / HTTPS / HTTP2,
Internet browsers
Server's responses and be able to analyze and verify them,
Google penalties for violating the guidelines and know the differences between them,
Important updates of the Google
Google's ranking factors (the most important ones) and search engine guidelines
How it all works keywords in the search engine,
Google courses (GAIQ, Ads)
Office suite - Excel, Docs
Content marketing
On-page SEO
HTML
CSS
JavaScript (Basic)
Google Analytics (Basic)
Google Tag Manager (Basic)
Google My Business
SEO tools support
Navigating the most popular CMSs, eg Wordpress
The ability to speed up page loading, reduce the weight of photos, how the browser cache works, what is css sprite, interpret the results provided by these tools -keyword analysis using of free and paid tools

Skills at the Senior SEO Specialist level:

Making marketing strategy
ON-page SEO (advanced)
OFF-page SEO - Linkbuilding (advanced)
HTML
CSS / SASS
RWD
JavaScript
CORS
API
Knowledge of frameworks (AngularJS, Backbone, Bootstrap, -Foundation and Vue.js)
Accessibility
RegExp
PHP
Python
NPM / YARN
Github
UX and UI
Google Ads
Google Tag Manager
Google Analytics
Content marketing
Excel
SQL
Navigate through popular CMS sites and stores.
Keyword analysis
Conducting trainings

Everyday tasks in the SEO industry
The most popular tasks in the daily work of a specialist:

keyword analysis
creating SEO audits
contact with developers
preparation of work reports
creating a campaign strategy
contact with the customer
implementation of audit changes
marketing of the company I work for (e.g. I run a blog)
linkbuilding
supervision over copywriters
creating editorial recommendations
conducting internal training
creating offers for customers
team management
obtaining links (I run the entire process)
creating a back-end
conducting trainings for clients
writing SEO texts

Summary
Becoming an SEO specialist may seem difficult in part, and it takes a lot of work, time and effort, but if you follow the plan and gradually develop your skills, it's something anyone can do.
Working in the SEO industry is both rewarding and creative.
It is also one of the jobs that is in demand as more and more companies are investing in online marketing ( and the pandemic has shown how important the e-commers segment is ). The most important thing in working with SEO is to be patient. SEO is a very slow process, so if you are a beginner it will take a while to learn the theory and a lot more time to gain hands-on experience.

Start today ❗️❗️ 🤓

In my next posts, I will say something more about individual skills or i'll say more about job as a Google Ads Specialist - something you can read here - https://afterweb.pl/google-ads-adwords/