Tag: Search Console

  • The questions your buyers ask AI, and how to answer them on your website

    The questions your buyers ask AI, and how to answer them on your website

    Your buyers are asking AI assistants whole questions about you, and the assistants are answering from whatever your website gives them.

    Fewer buyers type three words into a search box these days, and more of them put whole sentences, pasted briefs and rambling voice notes to an AI assistant, which draws on your website when it answers. Keyword tools are built around search volume, and long, specific questions barely register in them: almost 93% of the keywords in Ahrefs’ US database get fewer than ten searches a month (Ahrefs, 2026). A content plan built only from those tools can miss the very questions your buyers are asking. Here is how to find the questions your buyers ask AI, decide which ones deserve a page, and write the answer-first content that gets you cited.

    We start with why the questions reaching B2B websites have changed, then show what happened when we answered a set of them for a client. The second half is the method: where to find your own buyers’ questions, which ones deserve a page, and how to write answers a machine can quote and a buyer can trust. After reading, try writing a short list of the questions your website should be answering, then use the method to answer them. It works whichever assistant your buyers use, so it will still hold as ChatGPT, Gemini and Google’s AI Mode change.

    Why the questions got longer, and why short keywords no longer tell the whole story

    For most of the last twenty years websites were written to match short keywords, because that was what people typed. Many of the queries reaching B2B sites now are sentences, and a good share of them were typed into an AI assistant such as ChatGPT or Gemini, or sent to one as a voice note, rather than entered in a search box. Google vice-president of product management Hema Budaraju told AAP that early testers of AI Mode were asking queries “almost two to three times the length of traditional searches” (AAP, 2025). Once a buyer types a paragraph, AI Mode can split it into subtopics and search each one at the same time, a technique Google calls query fan-out (Google, 2025), looking across the pages it can index and display for an answer. Google’s Search Console documentation also treats a follow-up question in AI Mode as a new query (Google Search Console Help, 2025), which helps explain why site owners have started to see short replies such as “yes, pricing” in their performance reports, alongside long conversational queries that most likely come from AI Overviews or AI Mode (Search Engine Watch, 2026).

    We started seeing this in our own research some time before Search Console began showing it. Around the middle of 2024, while researching content for clients in complex sectors, we kept finding that the questions buyers asked drew almost no volume in keyword tools, while the one-to-three-word phrases those tools favour were held by the largest sites on the internet. Most short phrases in English have thousands of competing pages behind them, and a small group of very large sites takes most of the top positions. One analysis of 250,000 English search results found that 2.8% of ranking domains held three-quarters of the number-one spots (Detailed, 2025). Since mid-2024 our planning has been semantic, built around the topic, the questions people ask about it and the words they use when they ask. We still check exact-match terms, and they earn a place in a plan where they fit the topic.

    We see the same pattern in our own Search Console, where a growing share of the queries reaching contentifai.agency are full questions about AI crawlers, editing AI-written content and citation gaps, and the number of long queries has risen sharply since the spring. If you want to know how to show up in AI search, this is the part worth understanding. A buyer asks a long question, the assistant goes looking for a page that answers it, and whether your page gets used depends on whether it answers that question in plain terms. Our post on how to track AI traffic covers the measurement side.

    What happened when we answered a client’s buyer questions in advance

    Once we understood what was happening we wanted to test it on a real site, and the result is why answering buyer questions in advance is now a standard part of what we do for clients. Earlier this year we took a client in a specialist B2B sector and wrote the questions a prospect would put to an AI assistant before shortlisting them: can they do a specific thing, where they will and will not work, how they compare with the firms a buyer would shortlist alongside them, and what people say about them. Then we ran those questions through the assistants and read what came back.

    The site answered some of them well, some vaguely and a few not at all, so we wrote the answers, in direct language, on the pages a machine is most likely to check. From September, questions worded like these began appearing in the client’s Search Console as queries, about a dozen of them, with the answering pages at position one or two for nearly every one. Across September and October they gathered close to six hundred impressions between them. The wording of those queries points to AI Overviews and AI Mode, so most of those impressions mean the client’s page was shown inside an AI answer to somebody working out who to shortlist.

    In plain terms, a dozen questions that buyers put to AI assistants are now answered from the client’s own pages, near the top of the results, at the moment a buyer is choosing a shortlist. Someone reading that answer inside the assistant sees the client named and what the client can do, so by the time they reach the website or send an email they can already know what the firm does. That is the visibility stage of content marketing working as intended, and answering buyers’ questions this way is one of the gaps we notice most often when we first look at a B2B website.

    Four places to find the questions your buyers ask AI

    A couple of hours with these four sources will give you more questions than you can answer this quarter, and three of them use information your business already holds.

    Search Console. Open the performance report, filter queries to seven words or more, then to the ones that begin with how, what, can, does, should or which. Sort by impressions. For each one, note the page it landed on and read that page’s first paragraph to see whether it answers the question in the first two sentences or leaves the buyer to hunt for it. The pages that fail that test are the first ones to rewrite, and they are your first month of answer-first content.

    Sales and discovery calls. Write down the three questions most prospects ask before they buy and the objection that comes after the price. Our own three are consistent: how we use AI in our work, what it costs, and what it will do for them. We answer the first in how we balance human expertise with AI, the second with the prices on our services page and a custom quote on a call, and the third through the three pillars we plan every campaign around: brand equity (being known), engagement (being liked) and acquisition (being trusted enough to buy from).

    The inbox. Search your sent mail for “can you also” and “do you work with”. If a client has had to ask you something by email, it is a fair bet their peers are asking an assistant the same thing before they get in touch.

    The assistants themselves. Ask ChatGPT, Gemini and Perplexity what they would want to know before recommending a firm like yours, then ask them what they say about you today. Run the same short list of prompts once a month in a logged-out or temporary chat, so your own history does not colour the answers. Our AI readiness check is built on this, and for most clients the part they find most useful is simply hearing what the assistants say about them today.

    Which buyer questions deserve a page

    Buyer questions fall into four kinds, and sorting yours into them is the quickest way to decide what to write first.

    1. Capability questions ask whether you can do a specific thing: “can you handle a project of this size”, “do you work with firms outside the UK”.
    2. Comparison questions put you next to a named rival for a named job: “you versus them for this kind of work”.
    3. Trust questions ask what other people say about you: reviews, named clients, how long you have been doing it.
    4. Process questions ask how it works, how long it takes and what it costs.

    A question earns a page, or a section of one, when it recurs, when it is asked before purchase, and when a straightforward answer could make a buyer choose you or rule you out. Six questions answered properly will do more for you than a forty-question FAQ written to cover every possibility.

    Across clients we keep finding that the questions buyers ask are the ones a firm would rather not answer in public, and price is the most obvious example. “What happens if it goes wrong” and “why are you dearer than the firm down the road” are two more. These questions get answered on every sales call, so writing them down feels repetitive to the people who answer them, and because every rival faces the same questions, the first draft of the answer tends to sound like everyone else’s. There is also a natural reluctance to commit, because a cautious, PR-neutral answer feels safer than a specific one, and the result is a page that reads like every other firm’s. What we are trying to find in each case is an answer that serves the buyer and the firm at the same time and will still be accurate in a year or two, because in sectors with long buying cycles a page that chases this month’s trend is out of date before the buyer who needed it arrives.

    How to answer buyers’ questions with answer-first content

    Before you write a single answer, check that a machine can read your site at all. We have found a homepage set to noindex, a robots.txt file blocking every crawler on the list, and stale or missing sitemaps on more sites than we expected. Google’s own guidance is that a page has to be indexed and eligible to appear in Search with a snippet before it can be shown in AI Overviews or AI Mode (Google Search Central, 2025), so there is no point adding anything to a page the crawlers cannot reach. While you are in Search Console, check the generative AI setting under Settings as well: every site is included in AI Overviews and AI Mode by default (Google Search Console Help, 2026), but a site that has been excluded there will not have its links or content shown in them. Our post on how AI crawlers read your website covers the crawler-access gaps, and a readiness check or a site audit finds them on your own site, so fix those before you write anything new.

    Once the site can be read, write each answer the way a buyer would ask the question. These are the things we check on every page.

    • Put the question in the heading, phrased the way a buyer would phrase it, and answer it in the first two sentences. Add the detail underneath for the reader who wants it.
    • State what you do and do not do as straightforward facts. A machine is less likely to quote a page that describes itself in adjectives, as our post on writing content specific enough to be quoted explains.
    • Answer comparison questions by saying what you are good at and when a different kind of supplier is the better fit, and leave the rival out of it. We have watched a buyer walk away from a firm because its site spoke badly of a competitor. If you do the job better, the evidence on the page should make that obvious without a comparison.
    • Put the evidence for trust questions where a buyer and a machine can both find it: named people, reviews, case studies on their own pages. Assistants also check what other sites say about you, so keep your reviews and directory listings consistent with your own pages, as our post on why no single source gets you chosen by AI explains.
    • Add FAQs for the readers who scan. Many readers skim the headings, jump to the part that interests them, and bookmark the page to come back to, and a short FAQ serves that reader well. FAQ schema describes the questions and answers on the page to a machine, and it helps most once the page is readable and the answers are real.
    • Write it the way you would say it on a call. An assistant reading a stiff, lawyered FAQ aloud can make you sound like the other firms it has quoted, and a reader who scans will feel the same.

    We can write your buyers’ answer pages for you

    This is the service behind our monthly blog content. Each month we take the real questions your buyers put to AI, found in your Search Console, your sales calls and the assistants themselves, and write the pages that answer them, two a month, with the readability check, research, drafting and house editing all included. See monthly blog content for what a month looks like and what it costs, then read on for what the reports will and will not tell you.

    What the long-query report leaves out

    An impression records that a page was shown inside an answer. Search Console may show a follow-up question a buyer typed, but it does not show the sub-questions the AI generated behind the answer, so the paragraph the buyer wrote around “does [your firm] handle projects under a certain size” stays invisible to you, and you will not know which of the three firms in that prompt you were being compared with. In our view there is no reliable ROI formula for AI search yet, so treat any precise figure with caution and read the impressions as what they are, which is evidence that your pages are being shown to buyers inside AI answers and the first stage before engagement and enquiry. What works, in our experience, is something small done regularly. Once a month, read the long queries, pick one and write the answer, and after a year you will have twelve pages that say what you do in the words your buyers use.

    Start with the three questions you hear on every call

    You now know why the questions reaching your website are longer and more specific than the keywords it was built for, where to find them without buying anything, how to sort them into the four kinds that deserve a page, and how to write an answer a machine can quote and a buyer can trust. The first step takes an afternoon. Write down the three questions you hear on every sales call and check whether your website answers each one in its first two sentences, and any that fail are your first pages.

    AI search changes month to month and some of the reporting described here will have moved by the time you read this, but a website that answers real questions plainly has held its value through every change we have seen so far, and we expect it to keep doing so.

    If you would like a second pair of eyes, our AI readiness check shows what the assistants say about you today, and our monthly blog content turns your buyers’ questions into pages, two a month. Contact us to get started and tell us the three questions your buyers ask most.

    Frequently asked questions: FAQ schema, answer-first pages and the questions buyers ask AI

    These are the questions we hear most often when clients first look at their long queries. The short answers are below, and the detail is in the sections above.

    How do I find the questions customers ask AI about my business?

    Start with Search Console: filter to queries of seven words or more, then to question words, and sort by impressions. Add the questions prospects ask on sales calls, the ones clients email you, and what ChatGPT, Gemini and Perplexity say when you ask them what they would want to know before recommending a firm like yours.

    Why does Search Console show long questions with impressions but no clicks?

    AI Mode can split a buyer’s prompt into subtopics and search each one, and a follow-up in the same conversation counts as a new query. When a page is shown inside the AI’s answer Google can record an impression, and the buyer often reads the answer there without visiting. An impression with no click can mean the page answered the question, so read it alongside the enquiries and visits that follow.

    Should a B2B website have an FAQ page or answer questions inside its service pages?

    Both, with service pages first. The service page should answer capability and process questions in its opening paragraphs, because that is one of the first pages a machine checks. A short FAQ on the same page serves readers who scan, and a separate FAQ page suits recurring process questions that do not belong to one service.

    Is FAQ schema still worth adding now Google has dropped FAQ rich results?

    FAQ schema describes the questions and answers on the page in a form machines can read, and it is still worth adding even though Google stopped showing FAQ rich results on 7 May 2026 (Google Search Central, 2026). It will not help a page appear in Google if that page is not indexed, is blocked by robots.txt or is missing from the sitemap, and it cannot improve a vague answer.

    How do I answer comparison questions without running down a competitor?

    Say what you are good at, say when a different kind of supplier is the better fit, and leave the rival out of it. Buyers read negative comments about competitors as a signal about the firm that wrote them. If you are better, let the evidence and a straightforward statement of what you do speak for you.

    How do I get my business recommended by ChatGPT for a buyer’s question?

    Make the page readable to crawlers, put the question in a heading in the buyer’s own words, answer it in the first two sentences with specific facts, and support it with named evidence. Assistants tend to favour pages that answer a question directly and verifiably over pages that describe a firm in adjectives.

  • How to Track AI Traffic in GA4: A 15-Minute Check (and What It Can’t See)

    How to Track AI Traffic in GA4: A 15-Minute Check (and What It Can’t See)

    If you have no idea whether AI tools are sending people to your website, start with a 15-minute check in Google Analytics 4, then fill the gaps it leaves with three sources GA4 can’t reach.

    Updated 11 November 2026.

    Clients have been asking us some version of this question for about a year, and the wording has moved from “is AI reading my site?” to “how do I know, and how do I track it?” That is a healthy change, because it means AI search has become a channel owners want to see in the monthly report, next to organic, paid and referral.

    This guide starts where most small B2B businesses already have data: GA4. It walks through the built-in AI Assistant channel, the reasons it can undercount, a custom channel group that recovers your history, and the things GA4 cannot see at all. It is the practical companion to our plain-English guide to AEO and our piece on how AI crawlers read your website. After reading, try blocking out a quarter of an hour before your next marketing meeting, opening GA4 and working through the four steps, so you walk in knowing whether AI is a channel worth watching for your business.

    How much AI traffic should a small B2B site expect?

    For almost every website we look at, sessions from AI assistants are still a small share of the whole. If you run ads, or have strong organic or social traffic, visits that begin in ChatGPT or Perplexity will sit low in your channel report. The owners who care most are looking for a new channel, or have watched organic search soften and want to know where the readers went. For them, a small channel that is growing is exactly the pattern worth tracking early.

    What AI traffic looks like on our own site. In the twelve months to September 2026, about one in twelve sessions on contentifai.agency arrived from an AI assistant. ChatGPT sent the most, then Perplexity and Gemini, with a handful from Claude, and around half were engaged sessions. We don’t forecast from a channel this size. We watch its share each month.

    The word “sessions” is doing careful work in that box, because most of what AI companies do to a website involves no person at all. Cloudflare’s analysis of AI bot activity across its network found that training accounted for nearly 80% of AI crawling by July 2025, with search indexing at about 17% and user-triggered requests at roughly 3% (Cloudflare, 2025). The traffic you can see in analytics is that last, thin category: a person asked a question, the tool fetched or cited your page, and the person clicked. Everything else is machines reading, and that belongs to the crawler conversation rather than this one.

    The 15-minute AI traffic check in GA4

    Back in 2025 our answer to “can this be tracked?” was yes, with some work, and a prediction that the big platforms would bring it in-house once they had settled the data streams and the privacy questions. Google has now done that in GA4, which makes it the right place to start. You need Viewer access to your GA4 property for steps 1, 2 and 4, and Editor access for step 3. The first two steps take five minutes, the channel group takes five and the exploration takes the last five.

    Step 1: Open the AI Assistant channel (2 minutes)

    Since 13 May 2026, GA4 has sorted visits from recognised AI tools into their own channel. Google’s release notes describe a dedicated way to measure traffic from popular AI assistants, and its default channel group definitions describe AI Assistant as visits from sources like ChatGPT, Gemini, DeepSeek, Copilot or Grok, with the medium set to “ai-assistant” (Google, 2026). Go to Reports, then Acquisition, then Traffic acquisition, set the date range to the last 90 days, and look for the AI Assistant row beside Organic Search and Referral.

    Step 2: Note the share, the engagement and the tools (3 minutes)

    Write down two numbers: AI Assistant sessions as a share of all sessions, and the channel’s engagement rate against Organic Search. Then add Session source as a secondary dimension, so each tool shows by name. If you want to track Grok, ChatGPT or Perplexity traffic separately, this is the view that splits them.

    Why GA4’s built-in AI channel can undercount

    Before you treat that number as the full picture, check it against four gaps we see on client sites and our own.

    • It starts in May 2026. On our own site, AI visits only began appearing under AI Assistant in June 2026, and earlier ones sit under Referral.
    • It depends on a referrer. Some AI apps open links without passing one, so those visits land in Direct, and no GA4 setting can recover them.
    • Google’s list can lag new tools. A newer assistant may not be on the list yet, and its visits stay in Referral until it is added.
    • It leaves out Google’s own AI answers. Google’s channel definitions exclude AI Overviews and AI Mode, so clicks from them count as Organic Search.

    Steps 3 and 4 close the first and third gaps. The other two need the sources further down.

    Step 3: Build a custom AI channel group with a regex (5 minutes)

    A custom channel group lets you define your own AI channel and, according to Google’s custom channel groups guidance, apply it to your reports retroactively (Google, 2026). That gives you the history the built-in channel lacks.

    1. Go to Admin, then Data display, then Channel groups, and select Create new channel group. GA4 copies the default group for you to edit.
    2. Name it “AI traffic check”, add a new channel called “AI tools”, and set the condition to Session source, “matches regex”, with the expression below.
    3. Drag “AI tools” above Referral. GA4 places each visit in the first channel it matches, so order matters.
    4. Save, return to Traffic acquisition and switch the report to your new channel group.
    .*(chatgpt|openai|perplexity|gemini\.google|bard\.google|copilot|claude\.ai|grok|deepseek|meta\.ai|mistral|you\.com|poe\.com).*

    Two details save a lot of head-scratching. Google’s regex guidance says GA4 matches the full value and is case sensitive by default, which is why the expression starts and ends with “.*” and stays in lower case (Google, 2026). Standard GA4 properties allow two custom channel groups, so name this one clearly. Check the source list in step 2 once a quarter and add any new tool to the expression.

    Step 4: Save an exploration for page-level detail (5 minutes)

    Go to Explore and start a Free form exploration. Add Session source and Landing page as dimensions, Sessions and Engaged sessions as metrics, and a filter on Session source that matches the same regex. The result shows which pages AI tools send people to, and once you save it, next month’s check takes a minute.

    What GA4 can’t see

    GA4 counts people who clicked through to your site with a referrer intact. Several kinds of AI activity never reach it, and knowing which ones keeps your numbers in proportion.

    • Clicks from AI Overviews and AI Mode, which arrive looking like ordinary Google searches.
    • Visits with no referrer, which land in Direct.
    • AI crawlers, which fetch pages without running your analytics tag, so they appear in server logs and never in GA4.

    The three questions no tool answers yet

    With GA4 and the sources below running, you will know that AI tools read your site and roughly how many people they send. Three things will still be missing, and no tool on the market reliably fills the gap.

    What the person asked. Search Console’s AI report holds back the queries, and GA4 never had them. You can see that someone came from ChatGPT; you cannot see whether they asked “best content agencies in the North East” or “is schema markup worth it”, which are very different visits.

    What the tool said about you. An impression or a referral tells you a link was shown. It does not tell you whether the answer was accurate, whether you were the recommendation or the fourth name in a list of six, or whether the tool used your page to back up a fact in an answer about something unrelated to your services. Being a credible reference is good for you, but it is a different thing from being chosen.

    What happened next. The jump from “read an AI answer” to “became a client” is the one every marketer wants to measure, and it is still invisible. Appearing in AI search, being recommended by it and being bought from are three separate steps.

    This is why we do not hand clients an ROI formula for AI search: a reliable one does not exist yet. What we can say is simpler. If your buyers are asking AI tools questions in your category, you will want to be in the answers, and the sensible measure for now is whether your share of that channel grows month on month.

    Beyond GA4: three sources that fill the gaps

    GA4 tells you who arrived. These three sources tell you where Google’s AI showed your pages, which bots came to read them and what your buyers say themselves.

    Search Console: the generative AI performance report

    Clicks from AI Overviews and AI Mode count as Google search visits in GA4, so analytics cannot tell you how often Google’s own AI used your pages. Search Console now can, up to a point. The generative AI performance report shows how many times links to your site appeared in AI Overviews and AI Mode, by page, country, device and date, and Google rolled it out to every site worldwide on 31 August 2026 (Google, 2026). You will find it under Performance. It gives impressions only, with no queries, clicks or position, so treat it as a presence measure for seeing which pages Google’s AI draws on.

    Server or CDN logs: which bots came, and why

    Analytics counts people, whereas logs count requests, and the logs are where you see which AI companies are reading your site and why. The major providers publish separate user agents for training crawlers, search-index crawlers and fetches made for a live user, and a CDN such as Cloudflare groups them for you. A quarterly look is enough. Our AI crawlers post covers the named bots.

    Your forms: ask people where they found you

    Every method above depends on a referrer surviving the journey, and many will not. The one source you control is the person. Add “an AI assistant (ChatGPT, Gemini, Perplexity or similar)” as an option to every “how did you hear about us?” field you have: the contact form, the discovery call notes, the onboarding questionnaire. Ask at the first opportunity, because the enquiry that never converts still tells you the channel is working.

    Want AI traffic measured for you each month?

    This is a Contentifai service. Measure is our monthly AI visibility and web performance report: we review your Google Analytics, Search Console and presence in AI answers, explain what changed in plain English, and give you three prioritised recommendations. From £200 a month on a 28-day rolling agreement.

    Four mistakes that skew AI traffic numbers

    Plenty of businesses will want to run this in-house, and that is a good choice when someone has the time to look at the numbers each month. The GA4 check takes a quarter of an hour, and the other three sources an afternoon at most. The harder part is reading the results without fooling yourself, and over the past year we have watched the same four mistakes make AI traffic look bigger, or smaller, than it is.

    • Counting crawler hits as visitors. Server logs record every request a bot makes, so a single training crawl can look like hundreds of “AI visits” if it is added to the analytics total.
    • Reading Search Console AI impressions as clicks. The report contains no clicks, so an impression count is a presence figure and nothing more.
    • Waiting until the channel is big enough to matter before tracking it. Leave it until then and you will have no baseline to compare against when it does matter.
    • Relying on the built-in AI Assistant channel alone. It misses history, stripped referrers and newer tools, and a form field would have caught what analytics dropped.

    Avoid those four and the numbers you collect will be modest but trustworthy, which is the right footing for a channel still finding its size.

    Five lines for your monthly AI traffic report

    Once the tracking is in place, keep the monthly view to five lines, whether you write the report for yourself or ask an agency to produce it for you.

    1. Share of sessions from AI tools, using your custom channel group, and whether it is rising or falling.
    2. Engaged-session rate for that channel against organic, because a small channel that engages well is worth more than its size suggests.
    3. AI impressions by page in Search Console, to see which content Google’s AI is drawing on.
    4. The count of “AI assistant” answers on your forms, your one source that does not depend on a referrer.
    5. The pages AI tools send people to, from your saved exploration, which tells you what to write more of.

    That last line is where measurement connects back to the work. Tracking is the easy part now that Google has built most of it. What earns the citations is still the content: specific, useful, written so a machine can quote it, on a site the crawlers can read. Measurement tells you whether that work is paying off, and it cannot do the work for you.


    Tracking AI traffic: what we covered

    This guide covered a 15-minute check in GA4, the four reasons its AI Assistant channel can undercount, a custom channel group and exploration that recover your history, and the three sources that fill the gaps GA4 leaves.

    Google has changed its AI reporting twice in 2026 already, and the menu paths above may move again, so treat the method as the durable part and the clicks as the part to check.

    If you would rather have this measured and explained each month, our AI visibility and web performance report does it from £200 a month, and Monthly Blog Content writes the pages AI tools cite. Contact us to get started.

    Frequently asked questions about tracking AI traffic

    The questions owners and marketing managers ask us most about measuring AI search.

    How do I track AI traffic in Google Analytics 4?

    Open Reports, then Acquisition, then Traffic acquisition, and look for the AI Assistant channel, which GA4 has assigned automatically since 13 May 2026. For history before then, and for newer tools, create a custom channel group with a regex on Session source and apply it to your reports.

    Why does GA4 show so little AI traffic?

    Partly because the channel is still small for most sites, and partly because GA4 undercounts it: the AI Assistant channel starts in May 2026, some apps strip the referrer so visits land in Direct, and Google’s list can lag new tools.

    Can I track Grok, ChatGPT and Perplexity traffic separately?

    Yes. In the Traffic acquisition report, add Session source as a secondary dimension and each AI tool shows by name, for example chatgpt.com, perplexity.ai or grok.com. A saved exploration filtered with a regex keeps the view for next month.

    Does GA4 show traffic from Google AI Overviews and AI Mode?

    No. Google excludes AI Overviews and AI Mode from the AI Assistant channel, so those clicks appear as Organic Search. Search Console’s generative AI performance report shows how often your pages appeared in them, as impressions only.

    We have no idea if AI is sending us traffic. Where do we start?

    Start with GA4’s AI Assistant channel and note its share of sessions over the last 90 days. Add a custom channel group to recover earlier visits, then add an AI option to your “how did you hear about us?” form fields.

    What is the difference between an AI crawler visit and AI traffic?

    A crawler visit is a machine fetching your pages to train a model or build an index, and it shows in server logs, never in analytics. AI traffic is a person who asked an AI tool a question and clicked through, and it shows in GA4 as a session from an AI source.

GDPR Cookie Consent with Real Cookie Banner