Share

By Zak Ali

Okay, let me clarify: Your SEO playbook isn’t wrong…it’s just incomplete.

I’ve been working in organic search since 2016. I remember the golden age when a well-optimized page with decent backlinks could print traffic. I remember watching our rankings and traffic drop after the Helpful Content Update (HCU), and scrambling to follow Google’s guidance for recovery. And I still remember the process of coming to terms with the fact that search traffic wouldn’t be returning to pre-HCU levels.

Now, as we enter the LLM age, I’m seeing consultants and influencers pawn off ranking theories as fact and repackaging tired “SEO best practices” with buzzwords. Every week there’s a new silver bullet: “Chunk optimization!” “Schema markup is the answer!” “Bullets help you rank in ChatGPT!”

But beneath the noise, something fundamental has shifted. Instead of following every “hack of the week,” let’s acknowledge that no one has this figured out and shift to a longer-term strategy.

The Computational Realities Of LLMs

LLM-generated answers need to be trustworthy and valuable. But there’s a problem that every AI company faces and few SEOs acknowledge.

Running deep content analysis on every webpage on the internet for every query isn’t feasible. The computational cost and time required would be astronomical. So instead, LLMs use heuristics: educated shortcuts.

Take ChatGPT’s citations for instance. When it doesn’t use training data alone, it searches the web using existing search engines. What ranks on page one is what it scans to determine citations.

This is why people say good AI SEO is just good SEO. The distinction, however, is that where you rank in Google is seen more as a signal of your credibility for LLMs, not as the end game.

What Actually Drives Visibility

What really matters are authority and relevance, and how those definitions have evolved ever so slightly.

Authority

Traditional authority signals like getting cited by credible sites, demonstrating editorial standards and displaying trust signals like third-party reviews still work. Likewise, so do the entity connections between your business, authors and recognized authorities in your niche.

But what LLMs are doing is making authority become more real-time and contextual.

A site that published definitive content three years ago but hasn’t updated it? Less authoritative than a site actively engaging with current developments. An author with no digital footprint beyond their company blog? Less credible than one who shows up in industry discussions, podcasts and peer publications.

LLMs are pattern-matching machines. They’re looking for signals that you’re not just about a topic, you’re of it. Your brand should be woven into the conversations in your space.

This is where I feel SEO playbooks are falling short. People are still focused on optimizing for algorithms rather than optimizing for being genuinely known.

Relevance

In keyword-based search, relevance is straightforward. Does your page match the query terms?

In conversational search, relevance is contextual and cumulative. Consider the difference between these queries: “best credit cards” and “I have a 720 credit score, travel internationally twice a year for work, and want to consolidate my spending. What credit card should I get?”

The second query is specific, nuanced and assumes a back-and-forth dialogue. The user might follow up with: “What if I don’t want an annual fee?” or “How does that compare to others?”

So if search is becoming more conversational and less keyword-dependent, where does that leave keyword research? You still need to know what people are searching for. But keyword volume data won’t reveal the nuanced, long-tail conversational queries people ask LLMs.

Despite every chat-based search being unique and every answer personalized, your customers’ underlying needs remain constant. So if you need keyword data to tell you what those needs are, you have work to do.

No More Hacks

The problem with most “AI SEO” advice is that it’s trying to reverse-engineer LLM behaviour the same way we reverse-engineered Google’s. Find the pattern, exploit it, scale it.

But LLMs aren’t ranking algorithms in the traditional sense. They’re synthesis engines. They’re aggregating, weighing credibility signals, and constructing answers from multiple sources.

You can’t hack synthesis the way you could hack PageRank.

This is why chunking advice, schema recommendations, and formatting tips are marginal gains at best. They might help a little, but they won’t save a fundamentally weak content strategy.

The best strategy for SEO today is probably what we should’ve been doing all along.

A New SEO Playbook (That Isn’t Really New)

Get to know your customer on a deeper level. This is harder than plugging keywords into Ahrefs. It requires actually understanding your audience.

• Talk to your customers. Actually talk to them. What questions do they ask? What confuses them? What do they wish they knew before buying?

• Monitor communities. Places like Reddit, industry forums and comments sections can show you how real people articulate problems in their own words, not in “keywords.”

• Focus on intent, not keywords. Create content that addresses the full spectrum of questions someone might have around a topic, not just the highest-volume search term.

Then, start building up your authority.

• Build genuine authority. Not through link schemes, but through consistently showing up as a credible voice in your industry. Publish on platforms beyond your blog. Engage with industry conversations. Develop real subject matter experts, not just “content writers.”

• Create comprehensively useful content. Not 500-word keyword-stuffed blog posts, but resources that actually answer the full scope of what someone needs to know. Content that gets naturally referenced because it’s the best explanation available.

• Use first-party data. Use actual customer insights, from sales calls, support tickets and user research to understand what they’re trying to accomplish. Then turn those insights actionable in your content.

• Stay relevant and current. Plainly said, stale, outdated content is actively harmful to your authority.

Final Thoughts

The new SEO is anti-SEO. Stop trying to hack your way to the top. Start trying to be genuinely helpful to the people who are looking for a solution to their problems. The advantage will increasingly go to companies that do the hard work of building real expertise and communicating it clearly.

Feature image credit: Getty

By Zak Ali

Zak Ali is General Manager for Finder.com, a global financial comparison platform helping millions make better financial decisions. Read Zak Ali’s full executive profile here. Find Zak Ali on LinkedIn and X. Visit Zak’s website.

Sourced from Forbes

Write A Comment