All posts

GEO vs SEO: getting found when nobody googles anymore

Published July 28, 2026 · by Majd Ghithan, written with the help of AI

GEO vs SEO: getting found when nobody googles anymore

Think about the last time you needed to know something technical. Did you open Google, scan ten blue links, and click through? Or did you just ask ChatGPT or Claude and read the answer it gave you?

More and more, it's the second one. And that single change breaks the assumption the entire web was built on: that people search, see a list of links, and click yours. When someone asks an AI assistant directly, there is no list. There's one answer. If your content isn't inside that answer, you don't exist for that person — no impression, no click, nothing.

That's the gap GEO fills.

Two different games

SEO — Search Engine Optimization — is the game you know. You optimize so Google ranks your page, and you win a click when someone picks your blue link off the results page.

GEO — Generative Engine Optimization — is optimizing so AI assistants (ChatGPT, Claude, Perplexity, Gemini) cite or recommend you inside their answer. You don't win a rank. You win a mention.

They sound similar, and they overlap a lot. But the thing you're optimizing for is different:

SEOGEO
SurfaceGoogle results pageThe AI's answer itself
You win whenYour link ranks highThe model quotes or names you
The currencyClicks / trafficCitations / mentions
The readerSees ten options, picks oneSees one synthesized answer

Why GEO matters now, not later

The behavior already shifted. A developer debugging a Laravel queue at 1am doesn't open ten Stack Overflow tabs anymore — they paste the error into an AI and get a synthesized answer, often built from a handful of sources the model trusts. The question that decides your visibility is no longer "do I rank on page one?" It's "is my content one of the sources the model pulls from?"

What GEO actually asks of you

The good news: GEO isn't a separate stack you bolt on. It's mostly the same fundamentals as good SEO, aimed at a different target.

1. Let the AI crawlers in. This is the one people get backwards. AI assistants read the web through their own bots — GPTBot, ClaudeBot, PerplexityBot, Google-Extended. If your robots.txt blocks them, you've opted out of every AI answer on purpose.

Invisible to AI — blocks the crawlers
# robots.txt
User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

# congrats, you'll
# never be cited
Citable — lets them read you
# robots.txt
User-agent: *
Allow: /

Sitemap: https://you.com/sitemap.xml

# every crawler,
# including AI, welcome

2. Be extractable. LLMs pull claims out of your text. Clear headings, direct definitions, real numbers with sources, short declarative sentences — they all make your content easy to lift verbatim. A wall of vague prose gets skipped for the page that says the thing plainly.

3. Give machines structure. JSON-LD structured data (Article, BlogPosting, FAQPage), semantic HTML, and a clean sitemap tell both Google and the models exactly what your page is, who wrote it, and when. It's the difference between "some text" and "a dated, authored article about X."

4. Build authority signals. Models weight sources they can trust — a named author with a real bio and profile, links from other sites, being talked about elsewhere. GEO is partly won off your own site, in what the rest of the web says about you.

5. Consider an llms.txt. An emerging convention: a plain-text file at your root that points AI crawlers at your best, cleanest content — like a sitemap.xml written for models instead of search engines.

So is SEO dead?

No — and don't let anyone sell you that. Google still sends real traffic, and almost everything that helps GEO (crawlability, structure, clarity, authority) helps SEO too. Think of it as layers, not a replacement: SEO gets you into the index; GEO gets you into the answer. Do the shared fundamentals well and you're playing both games with one codebase. Skip GEO entirely and you're optimizing beautifully for a results page a growing share of your audience has stopped opening.

🤖 Heads up: this post was drafted with AI. Spot something wrong or off?Edit it on GitHub & open a PR
Majd Ghithan

Majd Ghithan

Full-Stack Engineer & Tech Lead

More posts

Eight months, no plan, nothing shipped. We shipped it in two.

Eight months, no plan, nothing shipped. We shipped it in two.

A team had been building a health platform for eight, maybe nine months. The client had run out of money. And there was nothing to ship — not because the code didn't exist, but…

July 28, 2026Read more
Laravel finally resizes images for you

Laravel finally resizes images for you

Every Laravel app I've built that takes an avatar ended the same way: composer require intervention/image, wire up a facade, write a little service class, and hope the next dev…

July 26, 2026Read more
What actually breaks at 100,000 users (that never breaks in a demo)

What actually breaks at 100,000 users (that never breaks in a demo)

The first time I shipped a dashboard to a hundred thousand active users, nothing I had tested was what broke. Everything that failed had passed every check I ran. It worked on m…

July 25, 2026Read more
Where your queue jobs go to die under load

Where your queue jobs go to die under load

The support ticket said "I never got my invoice email." I checked the logs. The job ran. It succeeded. failedjobs was empty. Everything said the email went out. It did not.

July 18, 2026Read more
The indexes you're missing (and the one that's hurting you)

The indexes you're missing (and the one that's hurting you)

A client sent me a query that took 4.2 seconds. One WHERE, one ORDER BY, a table with three million rows. I added a single index and it dropped to 11 milliseconds. They asked if…

July 11, 2026Read more
Caching in Laravel without the stampede

Caching in Laravel without the stampede

We cached the homepage's "trending products" query for five minutes. It was our slowest query — about 900ms, a big aggregation across orders. Caching it took the homepage from s…

July 4, 2026Read more
Taking one endpoint from 800ms to 80ms

Taking one endpoint from 800ms to 80ms

The order-details endpoint took 800 milliseconds. Not broken, just slow enough that the app felt heavy everywhere it was used. The team's instinct was "the server needs more mem…

June 27, 2026Read more
The N+1 you can't see (it's hiding in your accessors)

The N+1 you can't see (it's hiding in your accessors)

I've fixed hundreds of N+1 queries. The easy ones are right there in the controller — a foreach with $order-customer inside it, obvious the moment you read the code. Those aren'…

June 20, 2026Read more
The Friday deploy that charged customers twice

The Friday deploy that charged customers twice

It was 4:40 on a Friday. The change was tiny — a one-line tweak to how we called the payment provider, plus a bump to the queue worker's timeout. Small, tested, reviewed. I depl…

June 13, 2026Read more
Code review people don't dread

Code review people don't dread

I once left forty-one comments on a junior's pull request. I was proud of it. Thorough, I told myself. The next day he barely made eye contact, and his next PR sat open for a we…

June 6, 2026Read more
Hiring a mid-level Laravel dev: the signals that actually matter

Hiring a mid-level Laravel dev: the signals that actually matter

The best hire I ever made failed my first question. I asked him to explain service containers and he stumbled, went quiet, then said "honestly I use them every day but I've neve…

May 30, 2026Read more
Saying no to a feature without being the 'no' person

Saying no to a feature without being the 'no' person

For about a year I was the engineer everyone learned to route around. Not because I was wrong, I was usually right, but because my answer to new ideas was a flat "no, that'll br…

May 23, 2026Read more
My first 90 days as a tech lead (and the habit I had to break)

My first 90 days as a tech lead (and the habit I had to break)

Three weeks into leading my first team, I stayed late to "help" by rewriting a junior's feature myself. It was faster my way. I pushed it, felt productive, went home. The next m…

May 16, 2026Read more
Breaking down the task that scares you

Breaking down the task that scares you

There's a specific kind of ticket that makes my stomach drop. Not the hard ones, hard is fine. It's the ambiguous ones. "Migrate billing to the new provider." "Add multi-tenancy…

May 9, 2026Read more
1:1s that aren't just status updates

1:1s that aren't just status updates

For my first few months as a lead, my 1:1s were thirty minutes of me asking "so, what are you working on?" and nodding at answers I already knew from standup. We both left sligh…

May 2, 2026Read more
Get the business logic out of your controllers

Get the business logic out of your controllers

I once opened a OrderController@store method that was 240 lines long. Validation, a discount calculation, three model writes, a Stripe charge, two emails, a Slack notification,…

April 25, 2026Read more
Form Requests are the most underrated thing in Laravel

Form Requests are the most underrated thing in Laravel

Most Laravel devs meet Form Requests once, in a tutorial, use them to hold a rules() array, and never look deeper. That's a shame, because a Form Request is the cleanest place i…

April 18, 2026Read more
Testing Laravel without mocking everything

Testing Laravel without mocking everything

I inherited a codebase once with 900 unit tests and no confidence. Every test mocked the repository, mocked the model, mocked the mailer, mocked the thing three layers down. The…

April 11, 2026Read more
Three Eloquent features that clean up your models

Three Eloquent features that clean up your models

The messiest Laravel model I ever wrote wasn't messy because of Eloquent. It was messy because I ignored the parts of Eloquent that exist specifically to keep it clean. The same…

April 4, 2026Read more
The migration that locked the table for eight minutes

The migration that locked the table for eight minutes

The deploy looked boring. One migration, adding a lastseenat column to the users table with a default of now(). I'd written a hundred like it. I ran it at 2pm on a Tuesday becau…

March 28, 2026Read more
Chasing a memory leak in a Laravel queue worker

Chasing a memory leak in a Laravel queue worker

The alert came in at 4am: one of our queue:work processes had been OOM-killed. The supervisor restarted it, it processed jobs for about forty minutes, and got killed again. It w…

March 21, 2026Read more