If you’ve been on LinkedIn lately, you’ve probably seen the same claim: “SEO is dead.”

This time, the panic is SEO and AI – Google’s AI Overviews and other generative results. It’s fair to wonder whether people will still click through, whether SEO is worth it, and whether your content strategy needs a full reset.

But we’ve heard this before. Whenever a new search technology arrives, the headlines say traditional SEO is finished. It never is. SEO doesn’t disappear, it evolves, gets harder to game, and becomes more focused on what real people need. AI is just the latest shift.

This article cuts through the noise: what’s actually changing, and what a sensible AI-era search strategy looks like.

A short trip down SEO memory lane

To understand what’s happening now, it’s helpful to remember we’ve been here before.

  • 2015 mobile-first: “Desktop is dead.” It wasn’t. Sites that adapted improved.

  • AMP: Pitched as the future. For most, it was a temporary detour.

  • Voice search: Big predictions, smaller reality. Useful, but not a replacement.

  • Core Web Vitals: Speed mattered more, but it became one factor among many.

Now it’s SEO and AI Overviews. New tech, same pattern. SEO hasn’t died through any of these changes,  it’s become more technical, more user-focused, and better at rewarding genuinely helpful organisations. AI Overviews are another step on that path, not the end of it.

What’s actually changing with AI in Search?

Google hasn’t suddenly “learned everything” and stopped needing websites. The shift is that search is getting better at summarising and combining information from multiple sources.

Previously, you might have typed a question, opened several different results, and then pieced together the answer yourself. AI Overviews and large language models (LLMs) similar generative features try to do more of that legwork for you.

But AI can’t invent expertise. It relies on clear, trustworthy, well-structured content. If your site delivers that, you’re still in the mix.

AI hasn’t killed search, it’s expanded it

A lot of the anxiety around AI in search comes from the belief that tools like ChatGPT or Gemini will replace traditional search behaviour entirely. That isn’t what the data is showing.

What we’re seeing instead is that people aren’t stopping their use of Google and Bing; they’re adding AI tools into the mix. AI assistants are great when you need to brainstorm ideas, draft something from scratch, or get help writing code. When people need to verify a fact, buy insurance, book a flight, or make a meaningful decision, they still turn to search engines and websites.

The journey has changed rather than disappeared. AI sits alongside search instead of replacing it. That’s important, because it means the fundamentals of search visibility are still relevant, they just now influence AI-generated answers as well.

The Great Decoupling

One of the biggest practical changes is something we’re starting to see inside Google Search Console: what’s often called “the great decoupling“. Historically, in tools like Google Search Console, your impressions and clicks would move together. Now, they are splitting apart: impressions are skyrocketing while clicks may decrease.

Don’t panic. This happens because your content is being read and served directly within AI overviews (counting as an impression), satisfying the user immediately without a click.

Think of AI as a filter. It handles the low-intent ‘browsing’ traffic, meaning the users who do click through to your site are highly qualified and ready to convert. You are essentially trading traffic volume for traffic value.

Why retrieval matters: how AI systems actually work

To get your content into these AI answers, you need to understand how they think. These models use a process called RAG (Retrieval Augmented Generation).

Think of an AI model like a student taking an exam:

  • Closed Book: Without access to live data, the AI relies only on its training (which cuts off at a certain date). If you ask it about something recent, it might guess, or ‘hallucinate’.
  • Open Book (RAG): RAG allows the AI to browse the live internet to find the correct answer before responding.

Your goal is to make sure your website is the textbook the AI pulls from during that ‘open book’ test. If your content isn’t optimised for that initial retrieval, you simply won’t be the answer.

Technical SEO, clear headings, sensible internal linking, structured data, good formatting: these aren’t nice-to-haves. They are now the basic price of entry if you want AI systems to reliably find and use your content.

Your digital footprint is bigger than your website

Another major shift is that AI tools aren’t just training on your website. They learn from the open web. That includes:

  • Forums and communities like Reddit and Quora, where people share honest experiences and practical advice. These are heavily cited because they’re rich in user-generated content and real-world problem-solving.
  • YouTube, the second-largest search engine, where AI systems can use video transcripts as a source of truth.
  • Social platforms, where they can see whether your brand appears to be active, credible and consistent.

From an AI point of view, your “site” now includes your entire digital footprint. It isn’t enough to have a tidy website if your brand looks ghosted everywhere else.

That doesn’t mean you need to be everywhere, all the time. It does mean you should be present in the places that matter for your audience and that your messaging, facts and positioning stay consistent across them.

Multimodal content: not just “nice to have” anymore

Search used to be mainly about text. Now it’s multimodal by default: text, audio, video and images are all fair game as sources of information. AI Overviews and other generative panels are increasingly keen on citing original media: original videos, original audio, original imagery rather than generic stock. That has a few practical implications:

  • If you have a podcast, transcribing it turns your audio into searchable text that AI can understand.
  • If you have strong first-party data, turning it into infographics or visual summaries gives AI another trustworthy, distinctly “yours” asset to draw on.
  • If most of your visuals are stock photos, you’re essentially invisible at an image level. AI struggles to get any unique signal from them.

The goal is to publish useful ideas in multiple formats so they can travel further and are all feeding the same core message into different parts of the ecosystem AI is learning from.

Authority and trust: your defence against hallucinations

The biggest weakness of AI is hallucination. To avoid lying, AI tools are programmed to be risk-averse, they prioritise sources with high E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness).

Building authority is your ‘moat’ against AI competition. You can strengthen this by:

  • Using Author Bios: Link your content to real experts with LinkedIn profiles so the AI can ‘connect the dots’ on their expertise.
  • Digital PR: Earning mentions on high-authority sites (like universities or news outlets) signals to the AI that you are a credible source.
  • Consistency matters: if your organisation’s name, address and other key details don’t match across your site, your Google Business Profile, directories and social platforms, you look confusing, which is the opposite of trustworthy.

The more consistent signals you send that you are who you say you are, and you know what you’re talking about, the more comfortable AI tools will be citing you instead of someone else.

Local, personal and fresh beats generic

AI is getting better at nuance so a blanket, one-size-fits-everyone content doesn’t work as well as it once seemed to. Highly specific, locally relevant, audience-aware content is far more powerful. If you can reasonably localise or adapt content for different regions, contexts or user segments, it’s worth doing. 

Freshness matters too but not in the “change the date on an old blog post and hope nobody notices” sense. AI systems are clever enough to spot what’s sometimes called “artificial refreshing”. They care about meaningful updates:

  • New data
  • New case studies
  • New examples

If your content reflects what’s really happening now, it’s much more attractive as a retrieval source.

New ways to measure success

With new behaviours come new metrics. Traditional measures like rankings and organic sessions are still useful, but they don’t tell the whole story.

You’ll hear more talk about things like:

  • How often your brand or content appears in AI answers
  • How often you’re mentioned compared with competitors within AI-generated results 
  • How frequently your site is directly cited or quoted by AI.
  • How much traffic is being driven from AI platforms themselves.

On the analytics side, there’s no neat “AI” traffic channel in GA4 yet. Grouping traffic from major AI referrers into its own channel usually requires some custom configuration but it’s worth doing if you want to understand how those platforms contribute over time.

What you can do in the next 30, 60, 90 days

None of this means you need to tear everything up and start again. It does mean you should be deliberate about where you focus your energy. A practical way to think about it is in phases:

  • In the next month or so, focus on retrieval. Make sure your site’s technical foundations are sound. Fix major crawl issues. Tighten your headings and structure. Use question-and-answer formats where it makes sense. Implement or improve structured data so search engines and AI can understand your content more easily.
  • Over the following weeks, work on feeding better training data into the ecosystem. Update key content with real, recent information. Turn important ideas into multiple formats: blogs, short videos, audio clips, visuals and get them out onto the channels where your audience actually spends time.
  • Over the longer term, concentrate on building authority. Invest in digital PR where you can earn high-quality mentions. Strengthen author bios and expert profiles. Clean up your brand’s presence so your core details are consistent everywhere. Proactively answer questions about your brand so others don’t do it for you.

The details will vary between organisations, but the direction of travel is the same: make your content easy to find, easy to trust and easy to use, for both humans and machines.

So, is SEO dead?

No. It’s just harder to fake.

The future belongs to organisations that publish authentic, human-helpful content on technically sound foundations, and that show up consistently across the wider web. 

If you’re not sure where to start, whether it’s auditing your retrieval readiness, turning your existing content into multimodal assets, or making sense of AI-era search metrics, we’re already helping clients work through it.

Keep calm and optimise on. And if you’d like a clear, no-drama view of what AI in search means for your organisation, we’re here to talk.

Related content

How can we help?