SEO after AI Overviews — what actually matters in 2026

If you had a blog post ranking in position 2 for a question-format keyword in 2023, that position may still technically exist in 2026 — but nobody’s clicking through to read it. The AI Overview box sits above everything, answers the question in three sentences, and the user closes the tab. The click never happens.

This is not a hypothetical. Click-through rate data from GSC accounts across a range of industries shows the same pattern: impressions holding steady or growing for informational queries, clicks declining. The page still ranks. The traffic is gone.

The response from a lot of SEO commentary has been “optimise for AI Overview citations” — the theory being that if Google pulls your content into the summary box, you get brand visibility even without the click. That’s true as far as it goes. But it misses the bigger strategic question, which is whether informational SEO was ever the right investment for most Australian SMBs to begin with.

What's Actually Changed in the SERPs

AI Overviews aren’t uniformly deployed. They show up most aggressively for informational, how-to, and definitional queries — the question-answer format that long-tail content strategy has been built around for a decade. “What is X”, “how to Y”, “best Z for this situation” — these are exactly the queries Google has decided it can answer itself.

Transactional and commercial-intent queries still look much closer to traditional SERPs. If someone searches “SEO agency Melbourne” or “Meta Ads management Brisbane”, they get a map pack, some ads, and organic results. No AI Overview. Google’s business model depends on those clicks happening.

Local intent queries are similar — “plumber Parramatta”, “accountant Gold Coast” — the SERP structure hasn’t fundamentally changed because Google needs the local services product to function.

What’s died, or is dying, is the middle layer: informational content produced at volume by businesses trying to attract top-of-funnel traffic. Not because that content is bad, but because Google has decided it can synthesise it.

Why Long-Tail Informational Is Mostly a Dead End Now

The long-tail informational model worked because Google had to rank someone’s article to answer the question. Now it doesn’t. It can pull from multiple sources, synthesise, and present a clean answer without sending traffic anywhere.

The businesses that got hurt worst are those who built content programmes specifically around keyword gap analysis — find questions with decent volume and low competition, write 1,200 words answering them, watch rankings accumulate. That was a legitimate and effective strategy for several years. It’s not anymore, at least not at the scale it used to be.

The exception: genuine primary research, proprietary data, and opinions that can’t be synthesised because they’re the original source. If you publish a survey of Australian small businesses on their marketing spend and AI reads it to build an Overview, you’re cited. That drives brand recognition and sometimes direct traffic. The content is genuinely valuable to Google’s system because it can’t be replicated from elsewhere.

But that’s a different content model from writing explanatory posts to capture keyword traffic. It’s content as authority signal, not content as traffic channel.

Three Areas Worth Investing In

Technical foundations. This sounds routine because it’s been said for fifteen years, but the fundamentals have become more important as the margin for error shrinks. Core Web Vitals, clean crawlability, correct canonical structure, schema markup — these aren’t ranking factors in the conventional sense but they determine whether Google’s systems can confidently understand and trust your pages. An account with poor technical hygiene gets deprioritised in AI Overview citations too, not just organic results.

For Australian businesses specifically: mobile performance matters more here than in most markets because mobile-first behaviour is higher. A page that loads slowly on a 4G connection in regional Queensland isn’t just a UX problem — it’s a signal problem.

Brand mentions and digital PR. This is where SEO and PR converge in a way that matters. Google’s E-E-A-T framework cares about entity recognition — whether Google’s knowledge graph understands that your business is a real, credible entity in your industry. Brand mentions across credible Australian publications, industry directories, podcast appearances, and business listings all contribute to that entity signal.

This isn’t about link building in the traditional backlink sense, though links still matter. It’s about the breadth of citation that makes Google confident your brand exists and is respected in context. A business that appears in Smart Company, the AFR, an industry association newsletter, and a few credible directories looks very different in the knowledge graph than one that only appears on its own website.

Conversion-grade landing pages for transactional intent. The SERPs that still drive clicks are transactional — people who have already formed intent and are comparing options. Ranking for “SEO services Sydney” or “Google Ads management Melbourne” still means something because those queries produce clicks.

The mistake is treating those landing pages like content pages. They need to do conversion work, not just rank. That means clear positioning, specific proof — results, case studies, client names where permitted — pricing signals, and a frictionless next step. A page that ranks for a high-intent commercial query and converts at 0.8% is not an SEO win. Rankings without conversion are vanity.

The Visibility-to-Conversion Ratio

The framing shift that matters: stop optimising for rankings in isolation and start optimising for the ratio of search visibility to actual leads or revenue.

For most Australian businesses with under 20 staff and a focused geographic or industry footprint, the highest-value SEO investments are: keeping the technical house in order so existing pages perform well, building brand authority signals that compound over time, and making sure every page that ranks for a commercial-intent query is doing serious conversion work.

A meaningful share of SEO budgets currently going toward informational content production should be audited. If you can’t draw a clear line from that content to customer enquiries or revenue, it’s maintaining traffic that was never converting.

A practical test: in your last 12 months of GSC data, filter to pages with more than 100 clicks. Sort by conversion rate, linked to GA4 if you have it set up. The pages doing actual conversion work will be obvious. So will the pages that have traffic but no business value.

Our [SEO service page](/services/seo/) covers how we structure SEO programmes for Australian businesses in this environment — what we focus on and, equally, what we’ve stopped doing. If your current SEO investment is heavy on content volume but light on commercial outcomes, it’s worth a conversation.

For businesses running paid alongside organic, the [PPC ROI Calculator](/ppc-roi-calculator/) is useful for modelling whether shifting some SEO budget toward paid search would improve the overall visibility-to-conversion equation.

This week: open GSC, pull click data for the last 12 months, and count how many of your top-50 ranking pages by impressions are also in your top-50 by clicks. That gap is the AI Overview problem made visible.