Where We Are: SEO, LLMs, and Learning Again

November 2022 was not the beginning of artificial intelligence.
But it was the moment language —our language— fully entered the interface.

Just over three years have passed since then. And while the noise has been constant, we are no longer in a phase of discovery or fascination. We are in a phase of coexistence.
AI is no longer a promise: it is an active player in the digital ecosystem.

This is not a post about the future.
It’s a post about the present we are already living in.

When SEO Changed Its Counterpart

For years, SEO was —simplifying a lot— a dialogue between humans and search engines.
We published content, engines indexed it, ranked it, and served it.
There were implicit rules, clear metrics, and a recognizable goal: the click (and yes, conversions, time on site, bounce rate…).

Today, that dialogue has a new intermediary.
Language models no longer just link information: they interpret it, synthesize it, and reformulate it.
In many cases, they respond directly without sending users anywhere.

SEO hasn’t died.
But it is no longer sufficient as we used to understand it.

Not because Google will disappear —it won’t— but because it is no longer the only reader.
We now write for systems that don’t browse, don’t scroll, and don’t “read” like humans do, yet strongly influence which brands, concepts, and sources become references.

Learning Again (Even After 20+ Years)

There is something uncomfortable about admitting you need to learn again after spending half your life doing the same thing.
But there is also something deeply healthy about it.

My professional path has always been tied to SEO. I’ve seen trends come and go, algorithms change, and “definitive solutions” become obsolete in months.

This was different.

There was no single epiphany. There was an accumulation of signals:
clickless answers, increasingly aggressive snippets, models that cited —or didn’t cite— sources, clients asking why traffic was dropping while their brand appeared in conversations.

The question stopped being which tactic should we apply and became a much more uncomfortable one:
do we truly understand the new playing field?

At Mindset Digital, we decided not to watch from the sidelines.
Not because we had certainty, but precisely because we didn’t.

From that process came learnings we’ve documented on the agency blog (linking a few here if you’re interested in the more practical side):

• Basic SEO for LLMs
• Advanced SEO strategies for LLMs
• LLMs don’t read pages: they reconstruct patterns
• The 10 essential principles to rank in the era of generative engines

Not as closed formulas, but as a shared learning journal.

SEO for LLMs: Cutting Through the Noise

Over time, the term “SEO for LLMs” started circulating… along with plenty of confusion.

It’s not about magical prompts.
It’s not about tricking AI.
It’s not about “hacking” answers.
And it’s definitely not a checklist.

What we’ve learned is that it’s far more about architecture, coherence, and real authority than about tricks.

Optimizing for language models means accepting uncomfortable truths:
that clicks won’t always happen,
that perfect attribution won’t always exist,
that impact may be diffuse, but cumulative.

And still, it’s worth it.

What We’ve Learned So Far

• Authority cannot be improvised.
• Mediocre content no longer survives.
• Brand matters again.
• Context outweighs isolated keywords.
• Long-term coherence beats tactical optimization.

Models learn from what exists.
If you are not clearly explained, well structured, and well positioned within the ecosystem, you don’t exist for them.

2026: Fewer Shortcuts, More Judgment

I don’t believe 2026 will bring a dramatic rupture.
It will be a year of consolidation.

More answers without visible sources.
More weight on brand history.
More value in conceptual clarity.
Less room for empty content.

Tools will keep improving.
But the real differentiator will be knowing what to ask for—and why.

Closing the Loop

I recently wrote here about chasing balance.
About doing so even knowing it can never be fully achieved.

This professional path connects deeply with that idea:
technology and judgment,
speed and pause,
continuous learning without losing critical thinking.

Learning again not because it’s fashionable, but because it’s the only way to remain useful.
As professionals.
And ultimately, as people.

This is where we are.