AI SEO keyword research is the practice of finding the commercial-intent phrases that Google's AI Overviews reward with clicks, then adding those phrases to pages you already rank with. Google now answers many informational queries itself, so the fix is competitor gap analysis plus page refreshes, not always new content.
What AI SEO keyword research fixes after AI Overviews
AI SEO keyword research fixes the gap between ranking and getting clicked. Since Google introduced AI Overviews, a page can appear at the top of results, have its information summarized, and still receive no visit because the overview omits the source link. The strategy here treats that lost traffic as recoverable: find the terms where links still flow, and where competitors earn them.
In the video this article is built from, SEO consultant Chris Palmer demonstrates the issue on a large U.S. home-buying website. In Semrush organic research, under SERP features, the site showed about 2,600 AI Overview terms linking to the domain against roughly 71,000 terms where the AI Overview displayed information with no link back. He describes that difference as Google siphoning traffic, in the same way People Also Ask and featured snippets did before it.
Google's own documentation on AI Overviews describes them as AI-generated summaries with links to learn more, but publishers have widely observed that many queries produce summaries without visible source links. The practical takeaway is to stop assuming ranking equals traffic, and start auditing which of your terms still convert visibility into visits.
The evidence for this pattern comes from the transcript speaker's own client work and tool walkthroughs, so treat the specific numbers as illustrative of his accounts rather than an industry-wide measurement. The workflow, however, is reproducible with any Semrush subscription.
How to find the AI Overview gap in Semrush organic research
You find the gap by running a competitor's root domain through Semrush's organic research report and reading the SERP features table. The video's method works like this:
- Enter a competitor's root domain, not a single page, so you capture the whole site's ranking profile.
- Open the organic research report, which shows non-paid traffic, and scroll to the SERP features breakdown.
- Compare the 'AI Overview' SERP features that link to the domain against those that do not link.
- Filter the keyword list toward commercial and transactional intent, skipping branded terms.
- Repeat for your top three to five competitors to collect phrase ideas across your service areas.
Palmer runs this exact process twice: once on a 'we buy houses California' competitor site and once on a large personal injury law firm site, where he reports about 81,000 non-linking AI Overview terms against roughly 30,000 linking terms. The point of the second pass is volume of ideas. Each competitor surfaces different phrases, and the ones appearing across several competitors are the strongest refresh candidates.
This deliberately replaces the older page-refresh habit of working from Google Search Console impressions. Impressions show what you already rank for; competitor AI Overview data shows what is actively being rewarded right now, including terms you may never have targeted.
Create new pages or refresh existing ones: which works better
Refreshing existing pages is usually the faster and better option, according to the speaker's experience, with new pages as the fallback for genuinely new topics. A new page needs writing, indexing, and links before it can compete, which stretches into months. An existing page already has authority, an index entry, and often a ranking position to build on.
The refresh is a form of keyword gap analysis. Once you identify the commercial-intent phrases competitors win AI Overview clicks on, you add those phrases naturally to your existing informational pages. The speaker's reasoning is that informational pages can absorb high-commercial language without cannibalizing your location and service pages, because AI Overviews tend to surface on informational queries while your service pages target transactional ones.
Google's SEO starter guide supports the underlying mechanics: pages should target distinct search intent, and internal links from informational content toward commercial pages help both. A page titled around an informational query keeps its title, URL, and H1, and simply gains supporting commercial phrasing plus internal links into the relevant service page.
New high-commercial pages make sense mainly for new sites that lack established pages to refresh, or when a blended-SERP opportunity (below) calls for a separate 'best of' page. Otherwise the speaker reports better results from consistent, small refreshes driven by competitor data than from batches of new content.
Which keywords deserve a page refresh
The keywords worth targeting are commercial or transactional phrases that AI Overviews currently answer with a link, especially those won by competitors directly above you. Not all ranking terms are equal: a purely informational phrase with no click path back is low value, while a commercial phrase feeding an AI Overview with a cited source can produce calls and leads.
The filter in the video is explicit. Within the Semrush keyword list, search commercial intent first, transactional second, and skip branded results. Look for examples like 'companies that buy houses for cash in' or service-plus-location patterns such as 'car accident lawyer' combined with a city. These phrases carry purchase intent even when they sit on an informational page.
Two cautions preserve the scope of the claim. First, the strategy comes from one practitioner reporting his own client results, so it is experience-based evidence rather than a controlled study. Second, AI Overview linking behavior changes; a term that links today may stop linking next quarter, which is why the refresh process should be repeated on a schedule rather than done once.
Blended SERPs: winning two positions for one topic
A blended SERP is a results page where Google ranks both an informational-style page and a commercial page for related versions of the same query. Palmer says he noticed this pattern early and names it directly: you can hold a 'car accident lawyer Houston' service page and separately rank a 'best car accident lawyers Houston' page, giving your site two results on the same page.
The 'best of' variant is traditionally an informational query, but Google increasingly shows commercial-style list pages there alongside traditional results. Palmer adds that these blended pages frequently attract AI Overviews as well, making them doubly valuable: one position for the classic result, one for the AI summary.
The execution is simple to describe. Create a separate page with a 'best [service] [location]' pattern, distinct from your existing service page, and target it with the same commercial-intent research. Done on the same root domain, the two pages serve different intents and avoid cannibalization, which is the same logic that lets informational pages carry commercial phrasing safely.
Frequently asked questions
- Does this strategy require Semrush specifically? No. Semrush is what the video demonstrates because its organic research report splits SERP features into linking and non-linking AI Overview data. Any tool that exposes SERP features per keyword, or manual SERP checks, can support the same research, though the workflow is fastest with a dedicated SEO platform.
- Will adding commercial keywords to informational pages hurt my service pages? The speaker argues it should not, because informational and transactional queries resolve to different result types, and AI Overviews mostly attach to informational queries. Link internally from the informational page to the service page rather than duplicating the service page's exact target phrase, which keeps intents separate.
- How often should page refreshes be run? The video frames refreshes as an ongoing habit: repeat the competitor research periodically, since AI Overview linking behavior shifts over time. The speaker contrasts this with one-off refreshes based on Search Console impressions, which go stale faster.
- Is this strategy proven across industries? The reported results come from the speaker's own accounts in home buying and personal injury law, competitive local service niches. Treat it as a practitioner-reported method worth testing, not an independently verified study, and expect outcomes to vary by niche and by how often AI Overviews appear for your queries.
- What should a new site do differently? New sites lack established pages to refresh, so the video's advice is to create new pages, prioritizing the commercial-intent phrases identified through competitor AI Overview research rather than guessing at topics.
Turn your own SEO walkthroughs into written guides
The strategy above works because someone recorded a screen, explained a process, and published it as a video. If you explain SEO tactics, client work, or research methods on YouTube the same way, that knowledge is currently locked in a format searchers can't skim.
Skala blog turns a YouTube video into a structured written article: paste the video URL, get a transcription, and generate an editable draft you can review and publish. It suits exactly this kind of process-driven content, where the steps matter more than the production value of the video.
Fork this article
Start a new branch from the same video, shaped your way. You keep the credit; the original keeps the attribution.
A fork in another language is filed as a translation of this article, so the two pages point at each other. You can unlink it later from the editor.
0/240
You are creating
- Format
- For
- Language
- Source
- Your angle
You will be asked to sign in before it is generated.
Buy credits