The original is one click away. Open original ↗
Keyword research for Google and AI search in 2026
Executive overview
AI overviews now absorb 35% of clicks even when you rank number one. Informational keywords that once drove steady traffic are being answered directly by Google — no click required.
The fix is a five-part strategy: use a structured AI prompt to generate seed keywords and modifiers, multiply them through a keyword tool, vet each candidate with the BID method, prioritise tool-based keywords AI can't yet replace, and build brand presence so AI systems recommend you by name.
The new game is optimising for two systems simultaneously: Google search and AI search.
Generating seed keywords with the hero prompt
- Give your AI assistant context: site type, monetisation model, and target audience
- Ask for seed keywords (1–2 words, niche-specific) and modifiers separately — they must not share the same words
- Modifiers surface content format intent: best, how-to, calculator, checker, review
- Paste seeds into a keyword research tool's matching terms report, then filter by modifiers to generate hundreds of real keyword ideas
Vetting keywords with the BID method
BID = Business potential, Intent, Difficulty. A keyword must pass all three.
- Business potential: ranking number one should move the needle toward your site's goals — high-volume low-intent keywords (e.g. "what is espresso") rarely convert
- Intent: Google the keyword and read the SERP — if top results are product pages, a blog post won't rank there
- Difficulty: check referring domains and domain rating of top-ranking pages, not just the KD score; look for gaps where low-DR pages rank in the top 10
- If you have Ahrefs, filter by max KD 20 and lowest DR to surface qualifying keywords in one step
- Final check: Google the keyword and ask whether the AI overview fully satisfies the query — if yes, it's a traffic trap
Tool keywords as an AI-proof category
- Informational blog posts lost ~77% of traffic as AI overviews expanded, even with stable rankings
- Free tools (backlink checkers, mortgage calculators, word counters) have seen flat traffic — no AI overview appears for tool queries
- Find opportunities: use seeds like "coffee" or "espresso" in Keywords Explorer, then filter by modifiers like calculator, checker, generator
- One well-built tool can outperform 50 blog posts in both traffic and backlinks
Building brand presence for AI recommendations
- AI systems (Google, ChatGPT) are trained on public conversations — Reddit, Twitter, YouTube comments
- Brand mentions in those conversations directly influence which brands AI recommends in overviews and chat responses
- The Electric bike company gained AI overview citations after a YouTube sponsorship sparked widespread organic discussion
- To find gaps: use a brand radar tool, enter your brand and competitors, filter for prompts where competitors appear in AI results and you do not
- Shift the keyword research question from "what should I rank for?" to "what queries should my brand be associated with in AI search?"
More like this — when you're ready for early access.
Join the waitlist for a personal account and content recommendations based on what you're working on.
No spam. Unsubscribe at any time.
You're on the list. We'll be in touch before launch.