How AI Search Killed Keywords (And What to Target Now)

If your AI search SEO strategy is still built around a spreadsheet of keywords and monthly search volumes, you are optimising for a system that no longer exists in the same form.

Google’s AI Overviews now appear above organic results for a large and growing share of queries. Bing’s Copilot answers questions before users reach the results. ChatGPT and Perplexity cite sources directly. The traffic that keyword rankings once guaranteed is being intercepted before it ever reaches your site.

This is not a reason to panic — but it is a reason to change your approach.

What AI Overviews Actually Changed for Search Intent SEO

The core shift is this: Google no longer primarily needs to send users to a source. For informational queries — “how does [X] work”, “what is the best [Y]”, “explain [Z]” — it can now synthesise an answer directly. Users get what they came for without clicking.

This changes the calculus of search intent SEO entirely. The keywords that were easiest to rank for — broad informational terms with clear dictionary-style answers — are now the least valuable. If someone types “what is a conversion rate” and an AI Overview answers it perfectly in the results, the organic click rate for that query may fall 40–60%. Your position 1 ranking becomes worth a fraction of what it was two years ago.

The queries that still drive strong click-through rates are the ones where the AI Overview cannot fully substitute for a deeper source. Complex, nuanced, or high-stakes queries. Queries with strong commercial intent. Queries that require trust, comparison, or a specific context.

Understanding this distinction is the foundation of generative search optimisation.

The Queries AI Cannot Fully Answer

To win in AI search, you need to identify which queries in your space are AI-resistant. These tend to share a few characteristics:

Commercial intent with specificity. Queries like “best SEO agency for eCommerce in Melbourne” or “Google Ads management for service businesses” require a real decision. AI can summarise the category, but it cannot replace the evaluation process. These users are clicking.

Local and hyperlocal queries. “Plumber Collingwood” or “SEO consultant Perth” require local knowledge and local trust. AI Overviews may appear but organic and Maps results still receive strong engagement because proximity and reviews are decision factors the AI cannot fully resolve.

Brand and comparison queries. “NikoSEO vs [competitor]” or “reviews for [specific agency]” require real-world data and community trust. These remain high-click queries.

Process-heavy how-to content. Not all informational content is killed by AI. Long, step-by-step guides where the user needs to follow along, reference back, or save the content still perform — especially when they include original frameworks, screenshots, or specific numbers the AI cannot fabricate.

SGE Ranking and the Concept of Intent Clusters

The most useful shift in SGE ranking strategy is moving from individual keywords to intent clusters.

An intent cluster is a group of queries that share the same underlying job-to-be-done, even if the surface-level wording differs. “How to reduce Google Ads cost per lead”, “Google Ads CPC too high”, and “Google Ads wasted spend” are all expressions of the same intent: a business owner is burning money on clicks and wants to stop.

When you build content around intent clusters rather than isolated keywords, you do three things:

1. You create resources comprehensive enough to satisfy the full range of related queries — the kind of depth AI Overviews cite as sources rather than replace.
2. You signal topical authority to Google across a cluster of related terms, which strengthens your visibility for the whole group.
3. You write for the outcome the user actually wants, not the narrow keyword they happened to type.

How to Map Intent Clusters for Your Business

Start with your customers, not a keyword tool. Talk to your sales team. Review enquiries. Look at the questions that appear in your inboxes. Group them by the underlying problem, not the surface wording.

Then for each cluster, ask: does someone in this headspace need to click, or will an AI Overview satisfy them?

If the query is purely educational and the answer is simple, AI will intercept it. Don’t build content around it unless you can add something the AI cannot — original research, specific case data, a genuine process or framework from your own work.

If the query has any of these qualities — commercial intent, local specificity, comparison need, process depth, or trust requirement — it is worth targeting. Build a resource that earns the click by going deeper than anything the AI can synthesise.

Three Experiments to Run This Week

Audit your current top pages against AI Overviews. For each of your top 10 organic landing pages, search the query that drives traffic and check whether an AI Overview appears. If it does, assess whether the page is still generating clicks or whether traffic has quietly dropped. Prioritise rebuilding content that is being intercepted.

Identify your highest-intent commercial queries. List every service you offer. For each, write out three versions of the query a ready-to-buy customer might type. These are your most defensible ranking targets.

Test your content for citability. Search your primary topic area and read the AI Overview that appears. Is your content cited? If not, consider whether your content has the structure, specificity, and depth that makes it worth citing — clear headings, concrete numbers, named frameworks.

AI search SEO is not about fighting the algorithm. It is about making yourself indispensable in the moments when users still need to click.

Talk to us about building an SEO strategy for the AI search era →