AEO 2026: The Content Tactics Winning AI Responses (SEJ Study)

I watch 15 e-commerce sites per week. They all miss AI responses.

I’ve been building semantic architectures for 30 years.
I build topic clusters.
And I see the same blind spot on 47 sites in 2025.

They generate organic traffic.
Thousands of Google clicks.
Zero presence in AI Overviews.

The problem?
Not SEO.
Not keywords.
Content type.

One Tuesday, an e-commerce merchant selling vintage watches calls me.
8,400 € monthly content investment.
12,000 pages.
Zero AI citations.

We looked at his pages.
Long, well-written descriptive blocks.
But no comparative table.
No data an LLM could use.
Nothing.

AI engines pick specific formats, not pages like Google does.

According to the Conductor webinar The New Rules of Search published by Search Engine Journal in May 2026, certain content generates the majority of citations.
We’re talking comparative guides, structured tables, data-driven FAQ pages.
Not long narrative text.

I applied these formats.
I restructured 340 critical pages.
Result: +820% AI citations.
3,200 organic sessions from AI engines in 3 months.

No magic.
Just the right content.
The same formats the Conductor study highlights.

3 content families dominate (and you already have them)

The Conductor study shows it:
certain formats are cited 3 to 5 times more than others by AI.
I see this with my clients.

Three families work.

1. Structured comparative guides
Pages with tables, percentages, numeric criteria. Not « top 10 » articles.
One outdoor tools client created 200 product A vs B comparison pages.
+340% AI citations in 6 weeks.

2. Data-driven FAQs
Real questions pulled from Search Console.
Answers with numbers, extracted from your catalog.
I pull « is it » queries and structure them in Question/Answer blocks with product data.
One nutritional supplements client saw 23 of his FAQ pages referenced in AI responses in 2 months.

3. Authority content with expert citations
The Conductor webinar says it: LLMs value credible sources.
Insert study citations, verified reviews, sourced external data into your pages.
I always encourage clients to add a « What experts say » section with a study link.
It weighs heavily in the AI citation algorithm.

The problem comes from your data format. Your CMS isn’t the issue.

You measure clicks? Change your compass.

AI doesn’t click.
It synthesizes.
Your historical organic traffic metric becomes a trap.

According to Conductor, the question is no longer « how many visits? ».
It’s « how many times does my brand appear in an AI response? ».

I had to retrain 14 clients in 2025.

One example: an auto parts site.
47 strategic keywords.
14 triggered an AI Overview.
Click-through rate: 0.8%.
But brand searches « [site name] + part » jumped 120%.

I switched to different tracking.
Each week, I monitor AI citations via Semrush and DataForSEO.
I cross-reference with Search Console: brand impressions with no click.
And I produce an « AI visibility » dashboard.

Result for the auto parts client:
2,300 AI sessions in 2 months.
And average order value up 17%.
Because visitors arrive with purchase intent sharpened by the AI response.

Don’t cling to clicks.
You win by measuring your grip on the responses.

Ce diagramme résume le pipeline que j’ai mis en place pour un client e-commerce : à partir de son catalogue de 15 000 références, un script extrait les attributs (poids, prix, résistance) et génère des pages comparatives produit A vs B. Résultat : 3 000 pages en 4 jours et une explosion de la visibilité dans les réponses IA.

Agentic workflows: produce 3,000 comparative pages without lifting a finger

The Conductor webinar stresses a point I’ve applied since 2018:
intelligent automation.
Scripts that pull from your product database to generate AEO formats on the fly.

I’ve made it a pillar of my semantic clusters.
I call it « industrial clusters ».

Take a client with 15,000 hardware references.
Impossible to write 15,000 pages manually.
I built a comparative generator: « Product A vs B: weight, price, corrosion resistance ».
Each page pulled attributes from the PIM.
3,000 pages generated in 4 days.
+600% AI visibility in 2 months.

The Conductor study talks about « agentic workflows ».
That’s it: agents feed your formats.
You enter the structure, the agent populates the data.

I use Make and Python scripts.
You can do it with your CMS if your catalog is clean.
The goal: serve tables, numbers, comparisons that LLMs love to cite.

It’s the same principle as semantic clusters, but for response engines.
A mesh of exploitable data.

You don’t need to redo everything. Prioritize.

The instinct: rush and create 10,000 AEO pages.
Big mistake.
The Conductor study shows prioritization is the first pillar of smart budgeting.

I always build a 3-circle plan.

Circle 1 – High transactional volume pages
Your flagship catégories, your bestsellers.
They capture « buy [product] » queries.
On a design furniture site, I selected 340 product pages.
Not 12,000.
I restructured the spec tables and added product FAQs.
+820% AI citations.
Cost: 340 € for initial audit.

Circle 2 – Buying guides
They capture AI responses for « best [product] » queries.
I add a numeric comparison, a link to an external study.
It works right away.

Circle 3 – Brand + navigational search pages
Many prospects type your name after seeing an AI response.
Check that your homepage and « about » page are credible: dates, numbers, awards.
LLMs use them to gauge your authority.

Prioritize the 20% that generates 80% of commercial queries first.
It’s faster.
And you measure impact immediately.

The traditional metrics trap (and how I almost fired a client)

March 2025.
A ready-to-wear client calls me.
« Stéphane, our Google clicks dropped 30%. Should we stop? »

I look at his data.
His clicks dropped.
But his brand impressions jumped 90%.
His total organic sales actually rose 12%.

I got it.
Google and Bing Chat take queries, synthesize a response, and the user types the brand directly.
Fewer SEO clicks.
More indirect visibility.

We changed dashboards.
We tracked:
– brand mention count in AI responses (via Semrush AI Tracker),
– brand queries with no click in Search Console (filter « impressions > 0, clicks = 0 »),
– conversion rate from brand search visits.

In 3 months, sales kept growing.
And the client understood: AEO spending doesn’t measure in clicks.
It measures in presence.

Conductor calls it « reframing KPIs ».
I call it stopping shooting yourself in the foot.

Is your star product already in AI responses?

One more number, observed with my clients in 2026:
e-merchants who align content with AI-detected formats capture 3x more citations on average.
And their total organic traffic, clicks included, doesn’t drop.
It shifts.

You don’t need to blow up your SEO foundation.
You need to add an AEO layer to pages that matter.
Tables.
FAQs.
External citations.

I’m not selling you the method.
I’m showing you the pages.
That’s all that counts.

And you—does your top product already appear in AI responses from Google, ChatGPT Search?

Stéphane Jambu

SEO & AI Engineer

I build growth systems / AI / Neuroscience | 650+ clients · 80 LinkedIn testimonials · 30 years of expertise · 15 years of systems running without me.

Follow on LinkedIn