Berlins Today
Google Search Console now has an experimental AI tool that builds your performance report from a plain-English request. We used it to spot a topic our readers were already searching for and had no page about. We wrote the page, and it was picking up traffic within days.
Quick Summary - TLDR: The AI feature sits in the Performance report under Search results and turns a sentence into filters and comparisons It shows you what Google already associates your site with, which makes it a ready-made list of articles to write We found people reaching us for email verification searches with nothing relevant to read, published one guide, and it passed 1.2K views in two days You can repeat the same routine on any site in about half an hour What the Search Console AI feature does Google calls it AI-powered configuration, and it is marked as an experiment. It lives inside the Performance report for Search results. Instead of picking filters by hand, you describe the view you want, and it sets up the query, page, country, device, date range and comparison filters for you.
It then shows you what it applied so you can check it. It does not write your articles or tell you what to publish. It saves the clicking, and that matters because the real value in Search Console is in the filters most people never get round to setting.
How to use it, step by step Open Search Console and add your site. If you have not connected your website yet, start at Google Search Console and verify ownership. Data takes a little while to appear on a new property.
Go to Performance, then Search results. This is the report that shows clicks, impressions, click-through rate and average position. Click the small starry button.
It reads "Customise your performance reports using AI". A panel opens on the right called AI-powered configuration (Experiment). Type what you want to see.
Write it the way you would say it to a colleague. Read the suggested configuration. It lists the filters it plans to use.
For a request like "queries containing how to or what is" it builds a query filter using a regular expression, for example (how to|what is) . Click Apply. The table updates.
Change the period to the last day, 7 days, 28 days or 3 months, and click a column heading to sort by impressions or clicks. Better prompts to try Google offers a few starter prompts, such as how your traffic compares to last month. They are fine for a first look, but these get closer to ideas for new articles: "Show queries containing how to, best or cheap over the last 3 months." "Show queries that contain 2026 or 2027." "Queries containing free on mobile in the United Kingdom." "Compare the last 28 days with the previous 28 days for pages containing guide." "Show queries that contain a question word such as what, why, which or can." Then sort by impressions.
Searches that show your site often but earn few clicks are the interesting ones. Google already thinks you are relevant, and either your page is not a good match or you do not have one. Our case study We ran queries like these on our own property and noticed that people were seeing Startup Fortune in results for searches about email verification.
Phrases such as how to verify email addresses, how to find working email addresses and ways to verify emails in 2027 kept turning up. We had nothing that properly explained any of it. So we researched the subject with AI tools, wrote a full guide, and published it two days ago.
It is here: How to Verify Email Addresses and Choose the Best Email Verification Service in 2027 . In that short time it has already passed 1.2K views on our own counter, with no promotion beyond the page itself. One result does not prove a formula, and the traffic came from a mix of sources.
But the logic holds. We did not guess at a topic. We wrote about something people were already searching for and already linked to our site.
Turning the data into articles List the queries where you appear but have no dedicated page. Group them by intent, so a how-to, a comparison and a "best of" each get their own treatment. Write the page for the person searching, and answer the question in the first lines.
Check back in a few weeks and use the same report to see whether impressions and position moved. Our earlier posts on how to use long-tail keywords and how contextual long-tail keywords nearly doubled our search traffic explain how to shape the pages once you have the list. Will this get you cited by AI assistants?
Possibly, though nobody controls that. AI assistants tend to draw on pages that clearly answer a specific question, and the pages this routine produces are built around exactly that. Treat it as a good side effect, not a promise.
Frequently asked questions Is the AI feature free? Search Console itself is free, and the feature is part of it. It is still labelled an experiment, so it may change.
Do I need to know regular expressions? No. The tool writes them for you.
It helps to read the result so you know what is being matched. Why can't I see the button? Check that you are in Performance, then Search results, and that your site has data.
The feature has been rolling out in stages. How often should I do this? Once a month is plenty for most sites.
This article is posted in Guides News , check it out for more related stories.
Source: Startup Fortune
