In the old days of SEO, you could throw together a few keyword-heavy pages, buy some backlinks, and watch your rankings improve in what felt like real-time compared to the last ten years.

Well, welcome back - optimising for AI Overviews feels like we've all been dropped back into 2004.

The AI Overview gold rush

If you've spent ten minutes on Reddit or LinkedIn lately, you've probably seen people bragging about how easy it is to rank in Google's AI Overviews. They're not wrong, either. At least, not yet.

I've done a lot of testing, and the formula for gaining visibility in AI Overviews can be very simple:

  1. Write a handful of "Top 5 tools/services/products in 2025" posts. You can even get AI to write them for you.
  2. Include your company near the top and convey a positive sentiment.
  3. Repurpose them into LinkedIn articles, Reddit posts and comments, YouTube videos, some guest posts, and directory listings.
  4. Wait a few weeks...

Before long, you'll start showing up in AI-overviews across Google, and generative AI conversations in ChatGPT, Claude, and Perplexity.

It works, but only because the current AI retrieval systems are naïve.

Why it's working (for now)

AI Overviews (and other generative search tools) aren't ranking pages the old-fashioned way. They're summarising information from multiple sources - and those summaries rely heavily on:

  • Structured content (lists, comparisons, headings).
  • Consensus (how often entities appear across sources).
  • Recency (fresh, frequently updated content).

In other words: flood the web with consistent, well-formatted "Top X" content, and you'll look authoritative to AI; even if you're gaming the system.

To put it in numbers:

That means even small or mid-tier sites can punch way above their weight, for now.

It's early and messy, exactly like the Wild West of SEO circa 2004.

The hidden risks of optimising for AI: Overviews

Unfortunately, what's currently working might also cause organic performance issues elsewhere.

Google is huge and complex, and it can be slow to adapt when it releases new systems. But it is anything but stupid, and this won't last long.

The same thing happened with low-quality guest posting, and content farms, and many black-hat methods over the years - they worked beautifully until they didn't.

Here's where the current and manipulative AI Overview tactics could backfire:

Duplicate Content & Content Cannibalisation

Syndicating the same listicle across your blog, LinkedIn, guest posts and directories creates dozens of near-identical pages.

Search engines recognise repetition and either ignore duplicates or split signals across them, so it's likely that none of the pages will perform well. Chances are, LLMs will do the same soon, too.

Low-quality backlink profile from guest post spam

Obtaining numerous links from low-quality listicles results in backlinks from sites with duplicate content. That creates noisy link profiles that could potentially trigger quality filters or devalue your domain.

Manipulative syndication signals (fake consensus)

Reposting the same set of recommendations across multiple sites (often from the same network) creates a “consensus” that LLMs pick up. It’s an artificial signal that's easy for retrieval models to devalue once patterns are detected.

Spammy anchor text patterns

When every syndicated piece uses the same or similar anchor text links, it creates obvious manipulation patterns that can cause link-based algorithmic issues.

Triggering trust or quality downgrades

Google's "helpful content" systems (which now roll into the core algorithm) track patterns across an entire domain.

If your site suddenly resembles a "review mill" or a self-referencing comparison hub, that could be a red flag.

You might rank in AI Overviews, but lose visibility in organic SERPs, where 90%+ of search traffic still happens.

Creating brand fatigue

If you're everywhere, on every "Top X" list, every directory, every post - but always written by someone else, users clock it.

What was once clever looks manipulative and desperate. The trust you gain from AI visibility can evaporate when users realise it's engineered.

Misaligned authority signals

AI Overviews draw from multiple data points, but classic SEO signals - such as backlinks, brand mentions, and entity authority - still matter.

If all your new visibility is built on self-published content, it's a short-term lift - not an enduring signal of expertise.

How to optimise for AI Overviews without hurting your long-term SEO

AI Overviews will get harder to game. The question is: will you still be visible when that happens?

Here's how to future-proof your strategy:

Make content AI-readable, not AI-bait

  • Use clear headings, subtopics, and summaries so LLMs can extract context easily.
  • Don't over-optimise with repetitive phrasing or keyword stuffing.
  • Keep it useful enough that a human would still find it enjoyable to read.

Earn citations from others

It's one thing to say you're great. It's another when others say it for you.

Build real authority through PR, partnerships, and collaborations. AI systems value third-party reinforcement over self-promotion.

Diversify your signals

Generative search pulls from across the web - so make sure your brand appears in:

  • Author bios and structured data
  • LinkedIn articles and YouTube descriptions
  • Forums, Reddit, or Quora (authentic mentions help more than you think)

Track AI visibility properly

Use manual prompts or tools like Semrush's "AI Overview tracking" to spot when your content appears.

Note: impressions ≠ traffic. Many AI Overview clicks never reach your site - treat them as brand awareness, not conversions.

The future of AI overview SEO

It's tempting to jump on what works right now. But remember, Google giveth, and Google taketh away.

Optimising for AI Overviews today feels a lot like stuffing meta keywords in 2004. It's a short-term opportunity with risks, not a long-term strategy.

If you're smart, you'll use this period not to spam your way to the top, but to understand how AI interprets your brand.

Because when the dust settles - and it will - the winners will be the ones who built trust, not tricks.

Final thought

Yes, ranking in AI Overviews is fairly easy. But easy doesn't last.

So by all means, experiment. Just don't mistake early-access chaos for a sustainable strategy.