What Our AI Mode Person Habits Research Reveals About The Future Of Search

Our new usability research of 37 members throughout seven particular search duties clearly exhibits that individuals:

1. Learn AI Mode
2. Not often click on out, and
3. Solely depart when they’re able to transact.

From what we all know, there isn’t one other impartial usability research that has explored AI Mode to this depth.

In Could, I printed an intensive two-part research of AI Overviews (AIOs) with Amanda, Eric Van Buskirk, and his crew. Eric and I additionally collaborated on Propellic’s journey trade AI mode research.

We labored collectively once more to convey you this week’s Progress Memo: a research that gives essential insights and validation into the behaviors of individuals as they work together with Google’s AI Mode.

Since neither Google nor OpenAI (or anybody else) offers person knowledge for his or her AI (Search) merchandise, we’re filling a vital hole.

We captured display recordings and think-aloud classes by way of distant research. The 250 distinctive duties collected present a sturdy knowledge set for our evaluation. (The entire methodology is offered on the finish of this memo, together with particulars concerning the seven search duties.)

And also you may be shocked by a number of the findings. We have been.

It is a longer submit, so seize a drink and settle in.

Picture Credit score: Kevin Indig

Govt Abstract

Our new usability research of Google’s AI Mode reveals how profoundly this function adjustments person conduct.

• AI Mode holds consideration and retains customers inside. In roughly three‑quarters of the full person classes, customers by no means left the AI Mode pane – and 88 % of customers’ first interactions have been with the AI‑generated textual content. Engagement was excessive: The median time by job sort was roughly 52-77 seconds.
• Clicks are uncommon and principally transactional. The median variety of exterior clicks per job was zero. Yep. You learn that proper. Ze-ro. And 77.6% of classes had zero exterior visits.
• Individuals skim however nonetheless make choices in AI Mode. Over half of the duties have been categorised as “skimmed shortly,” the place customers look on the AI‑generated abstract, type an opinion, and transfer on.
• AI Mode delivers “web site varieties” that match intent. It’s not nearly assembly search question or immediate intents; AI Mode is citing sources that match particular web site classes (like marketplaces vs overview websites vs manufacturers).
• Visibility, not site visitors, is the rising forex. Members made their model judgments straight from AI Mode outputs.

TL;DR? These are the core findings from this research:

• AI Mode is sticky.
• Clicks are reserved for transactions.
• AI Mode matches web site sort with intent.
• Product previews act like mini product element pages (aka PDPs).

However earlier than we dig in, a fast shout-out right here to the crew behind this research.

Along with Eric Van Buskirk’s crew at Clickstream Options, I performed the primary broad usability research of Google’s AI Mode that uncovers not solely essential insights into how folks work together with the hybrid search/AI chat engine, but additionally what sorts of branded websites AI Mode surfaces and when.

I wish to spotlight that Eric Van Buskirk was the analysis director. Whereas we collaborated carefully on shaping the analysis questions, areas of focus, and methodology, Eric managed the crew, oversaw the research execution, and delivered the findings. Afterward, we labored aspect by aspect to interpret the info.

Click on knowledge is a good first cross for evaluation on what’s occurring in AI Mode, however with this usability research particularly, we basically seemed “over the shoulder” of real-life customers as they accomplished duties, which resulted in a sturdy assortment of knowledge to tug insights from.

Our testing platform was UXtweak.

Boost your skills with Growth Memo’s weekly expert insights. Subscribe for free!

Google’s own Sundar Pichai has been crystal clear: AI Mode isn’t a toy; it’s a proving ground for what the core search experience will look like in the future.

On the Lex Fridman podcast, Pichai said (bolding mine):

“Our current plan is AI Mode is going to be there as a separate tab for people who really want to experience that… But as features work, we’ll keep migrating it to the main page…” [1]

Google has argued these new AI-focused options are designed to level customers to the net, however in apply, our knowledge exhibits that customers stick round and make choices with out clicking out. In idea, this might not solely impression click-outs to natural outcomes and citations, but additionally cut back exterior clicks to adverts.

In August, I explored the truth behind Google’s personal product cannibalization with AI Mode and AIOs:

Proper now, in response to Similarweb knowledge, utilization of the AI Mode tab on Google.com within the US has barely dipped and now sits at simply over 1%.

Google AIOs at the moment are seen by greater than 1.5 billion searchers each month, and so they sit entrance and middle. However engagement is falling. Customers are spending much less time on Google and clicking much less pages.

However as Google rolls AI Mode out extra broadly, it brings the largest shift to Search (the largest buyer acquisition channel there may be) ever.

Conventional web optimization is very efficient within the new AI world, but when AI Mode actually turns into the default, there’s a probability we have to rethink our arsenal of techniques.

Making ready for the way forward for search means treating AI Mode because the vacation spot (not the doorway), and determining the best way to present up there in ways in which truly matter to actual person conduct.

With this research, I sought out to find and validate precise person behaviors inside the AI Mode expertise when enterprise quite a lot of duties with differing search intents.

1. AI Mode Is Sticky

Picture Credit score: Kevin Indig

Key Stats

Individuals learn first and often keep contained in the AI Mode expertise. Right here’s what we discovered:

• Nearly all of classes had zero exterior visits: that means, they didn’t depart AI Mode (in any respect).
• ~88% of customers’ first interplay* inside the function was with the AI Mode textual content.
• Typical person engagement inside AI Mode is roughly 50 to 80 seconds per job.

These three stats outline the AI Mode search floor: It holds consideration and resolves many duties with out sending site visitors.

*Right here’s what I imply by “interplay:”

• An “interplay” inside the person duties = the participant meaningfully engaged with AI Mode after it loaded.
• What counts as an interplay: Studying or scrolling the AI Mode physique for greater than a fast look, together with scanning a outcome block just like the Buying Pack or Proper Pane, opening a service provider card, clicking an inline hyperlink, hyperlink icon, or picture pack.
• What doesn’t depend as an interplay: Transient eye flicks, cursor passes, or hesitation earlier than participating.

Customers are in AI Mode to learn – not essentially to browse or search – with ~88% of classes interacting with the output’s textual content first and spending one minute or extra inside the AI Mode expertise.

Plus, it’s attention-grabbing to see that customers spend greater than double the time in AI Mode in comparison with AIOs.

The general engagement is way stronger.

Picture Credit score: Kevin Indig

Why It Issues

Deal with the AI Mode panel like the first studying floor, not a teaser for blue hyperlinks.

AI Mode is a contained expertise the place sending clicks to web sites is a low precedence and giving customers the most effective reply is the best one.

Because of this, it fully adjustments the worth chain for content material creators, firms, and publishers.

Perception

Why do different sources and/or AI Mode analysis analyses say that customers don’t return to the AI Mode function fairly often?

My idea right here is that, as a result of AI mode is a separate search expertise (at the very least, for now), it’s not as seen as AIOs.

As AI Mode adoption will increase with Google bringing Gemini (and AI Mode) into the browser, I anticipate our research findings to scale.

2. Clicks Are Reserved For Transactions

Whereas clicks are scarce, buy intent isn’t.

Members within the research solely clicked out when the duty demanded it (e.g., “put an merchandise in your buying cart”) or in the event that they browsed round a bit.

Nevertheless, the shopping clicks have been so few that we are able to safely assume AI Mode solely results in click-outs when customers wish to buy.

Even prompts with a comparability and informational intent are likely to hold customers contained in the function.

• Buying prompts like [canvas bag] and [tidy desk cables] drive the best AI Mode exit share.
• Comparability prompts like [Oura vs Apple Watch] present the bottom exit share of the duties.

When members have been inspired to take motion (“put an merchandise in your buying cart” or “discover a product”), the vast majority of clicks went to buying options like Buying Packs or Service provider Playing cards.

Picture Credit score: Kevin Indig

18% of exits have been brought on by customers exiting AI Mode and going straight to a different web site, making it a lot more durable to reverse engineer what drove these visits within the first place.

Research transcripts affirm that members usually share out loud that they’ll “go to the vendor’s web page,” or “discover the product on Amazon/ebay” for product searches.

Even when evaluating merchandise, whether or not software program or bodily items, customers barely click on out.

Picture Credit score: Kevin Indig

In plain phrases, AI mode eats up all TOFU and MOFU clicks. Customers uncover merchandise and type opinions about them in AI Mode.

Key Stats

• Out of 250 legitimate duties, the median variety of exterior clicks was zero!
• The immediate job of [canvas bag] had 44 exterior clicks, and [tidy desk cables] had 31 clicks, accounting for two-thirds of all exterior clicks on this research.
• Comparability duties like [Oura Ring vs Apple Watch] or [Ramp vs Brex] had only a few clicks (≤6 whole throughout all duties).

Right here’s what’s attention-grabbing…

Within the AIOs Overviews usability research, we discovered desktop customers click on out ~10.6% of the time in comparison with virtually 0% in AI Mode.

Nevertheless, AIOs have natural search outcomes and SERP Options under them. (Individuals click on out much less in AIOs, however they click on on natural outcomes and SERP options extra usually.)

Zero-Clicks

• AI Overviews: 93%*
• AI Mode: ~100%

*Understand that members of the AIO usability research clicked on common natural search outcomes. The 93% pertains to zero clicks inside the AI Overview.

On desktop, AI Mode produces roughly double the in-panel clickouts in comparison with the AIO panel. On AIO SERPs, whole clickouts can nonetheless occur by way of natural outcomes under the panel, so the page-level charge will sit between the AIO-panel determine and the basic baseline.

An essential notice right here from Eric Van Kirk, the director of this research: When evaluating the AI Mode and AI Overview research, we’re not precisely evaluating apples to apples. On this research, members got duties that will immediate them to go away AI Mode in 2/7 questions, and that accounts for almost all of outbound clicks (which have been fewer than three exterior clicks). Then again, for the AIO research, probably the most transactional query was “Discover a transportable charger for telephones beneath $15. Search as you sometimes would.” They weren’t instructed to “put it in a buying cart.” Nevertheless, the insights gathered concerning person conduct from this AI Mode research – and the sample that customers don’t really feel the necessity to click on out of AI Mode to make further choices – nonetheless stands as a stable discovering.

The larger image right here is that AIOs are like a truth sheet that steers customers to websites ultimately, however AI Mode is a closed expertise that not often has customers clicking out.

What makes AI Mode (and ChatGPT, by the best way) difficult is when customers abandon the expertise and go on to web sites. It messes with attribution fashions and our potential to grasp what influences conversions.

3. AI Mode Matches Web site Sort With Intent

Within the research, we assess what forms of websites AI Mode exhibits for our seven duties.

The categories are:

• Manufacturers: Sellers/distributors.
• Marketplaces: amazon.com, ebay.com, walmart.com, homedepot.com, bestbuy.com, goal.com, rei.com.
• Assessment websites: nerdwallet.com, pcmag.com, zdnet.com, nymag.com, usatoday.com, businessinsider.com.
• Publishers: nytimes.com, nbcnews.com, youtube.com, thespruce.com.
• Platform: Google.

Picture Credit score: Kevin Indig

Buying prompts path to product pages:

• Canvas Bag: 93% of exits go to Model + Market.
• Tidy desk cables: 68% go to Model + Market, with a visual Writer slice.

Comparisons path to critiques:

• Ramp vs Brex: 83% Assessment.
• Oura vs Apple Watch: cut up 50% Model and 50% Market.

When the person has to carry out a repute examine, the result’s cut up model and publishers:

• Liquid Dying: 56% Model, 44 % Writer.

Google itself exhibits up on buying duties:

• Retailer lookups to enterprise.google.com seem on Canvas Bag (7%) and Tidy desk cables (11%).

Try the top-clicked domains by job:

• Canvas Bag: llbean.com, ebay.com, rticoutdoors.com, enterprise.google.com.
• Tidy desk cables: walmart.com, amazon.com, homedepot.com.
• Subscription language apps vs free: pcmag.com, nytimes.com, usatoday.com.
• Bottled Water (Liquid Dying): reddit.com, liquiddeath.com, youtube.com.
• Ramp vs Brex: nerdwallet.com, kruzeconsulting.com, airwallex.com.
• Oura Ring 3 vs Apple Watch 9: ouraring.com, zdnet.com.
• VR arcade or good residence: sandboxvr.com, enterprise.google.com, yodobashi.com.

Firms want to grasp the taking part in subject. Whereas basic web optimization allowed mainly any web site to be seen for any person intent, AI Mode has strict guidelines:

• Manufacturers beat marketplaces when customers know what product they need.
• Marketplaces are most well-liked when choices are broad or generic.
• Assessment websites seem for comparisons.
• Opinions spotlight Reddit and publishers.
• Google itself is most seen for native intent, and generally buying.

As SEOs, we have to think about how Google classifies our web site primarily based on its web page templates, repute, and person engagement. However most significantly, we have to monitor prompts in AI Mode and have a look at the location combine to grasp the place we are able to play.

Websites can’t and received’t be seen for all sorts of queries in a subject anymore; you’ll must filter your technique by the intent that aligns along with your web site sort as a result of AI Mode solely exhibits sure websites (like overview websites or manufacturers) for particular forms of intent.

Product previews present up in about 25% of the AI Mode classes, get ~9 seconds of consideration, and other people often open just one.

Then? 45% cease there. Many opens are fast spec checks, not a clickout.

Picture Credit score: Kevin Indig

You possibly can simply see how some product suggestions by AI Mode and on-site experiences are fairly irritating to customers.

The post-click expertise is essential: basic finest practices like critiques have a big effect on making probably the most out of the few clicks we nonetheless get.

See this instance:

“It seems prefer it has plenty of optimistic critiques. That’s one factor I’d have a look at if I used to be going to purchase this bag. So this may be the one I’d select.”

In buying duties, we discovered that model websites take the vast majority of exits.

Compared duties, we found that overview websites dominate. For repute checks (like a immediate for [Liquid Death]), exits to manufacturers and publishers have been cut up.

• For transactional intent prompts: Manufacturers take in most exits when the duty is to purchase one merchandise now. [Canvas Bag] exhibits a robust tilt to model PDPs.
• For repute intent prompts: Model websites seem alongside publishers. A immediate for [Liquid Death] splits between liquiddeath.com and Reddit/YouTube/Eater.
• For comparability prompts: Manufacturers take a again seat. [Ramp vs Brex] exits go principally to overview websites like NerdWallet and Kruze.

Given customers can now straight checkout on ChatGPT and AI Mode, shopping-related duties may ship even fewer clicks out.^[2, 3]

Due to this fact, AI Mode turns into a totally closed expertise the place even buying intent is fulfilled proper within the app.

Clicks are scarce. Affect is plentiful.

The information provides us a actuality examine: If customers proceed to undertake the brand new means of Googling, AI Mode will reshape search conduct in methods SEOs can’t afford to disregard.

• Technique shifts from “get the press” to “earn the quotation.”
• Comparisons are for belief, not site visitors. They cut back exits as a result of customers really feel knowledgeable contained in the panel.
• Retailers ought to optimize for decisive exits. Give costs, availability, and proof above the fold to transform the few exits you do get.

You’ll must earn citations that reply the duty, then win the few, high-intent exits that stay.

However our research doesn’t finish right here.

Right this moment’s outcomes reveal core insights into how folks work together with AI Mode. We’ll unpack extra to contemplate with Half 2 dropping subsequent week.

However for many who like to dig into particulars, the methodology of the research is included under.

Methodology

Research Design And Goal

We performed a mixed-methods usability research to quantify how Google’s new AI Mode adjustments searcher conduct. Every participant accomplished seven reside Google search prompts by way of the AI Mode function. This design permits us to look at each the mechanics of interplay (scrolls, clicks, dwell, belief) and the qualitative reasoning members voiced whereas finishing duties.

The duties:

1. What do folks say about Liquid Dying, the beverage firm? Do their drinks enchantment to you?
2. Think about you’re going to purchase a sleep tracker and the one two accessible are the Oura Ring 3 or the Apple Watch 9. Which might you select, and why?
3. You’re getting insights concerning the perks of a Ramp bank card vs. a Brex Card for small companies. Which one appears higher? What would make a enterprise change from one other card: charge element, eligibility high-quality print, or rewards?
4. Within the “Ask something” field in AI Mode, enter “Assist me buy a water-proof canvas bag.” Choose one that most closely fits your wants and you’ll purchase (for instance, a digital camera bag, tote bag, duffel bag, and many others.).
   • Proceed to the vendor’s web page. Click on so as to add to the buying cart and full this job with out going additional.
5. Evaluate subscription language apps to free language apps. Would you pay, and in what state of affairs? Which product would you select?
6. Suppose you might be visiting a buddy in a big metropolis and wish to go to both: 1. A digital actuality arcade OR 2. A sensible residence showroom. What’s the title of the town you’re visiting?
7. 1. Suppose you’re employed at a small desk and your cables are a large number. 2. Within the “Ask something” field in AI Mode, enter: “The system cables are cluttering up my desk house. What can I purchase immediately to assist?” 3. Then select the one product you assume could be the most effective answer. Put it within the buying cart on the exterior web site and finish this job.

Thirty-seven English-speaking U.S. adults have been recruited by way of Prolific between Aug. 20 and Sept. 1, 2025 (together with members in a small group who did pilot research).*

Eligibility required a ≥ 95% Prolific approval charge, a Chromium-based browser, and a functioning microphone. Members visited AI Mode and carried out duties remotely by way of their desktop pc; invalid classes have been excluded for technical failure or non-compliance. The ultimate dataset accommodates over 250 legitimate job information throughout 37 members.

*Pilot research are performed first in distant usability testing to establish and repair technical points – like screen-sharing, job setup, or recording issues – earlier than the principle research begins. They assist refine job wording, timing, and directions to make sure members interpret them appropriately. Most significantly, pilot classes affirm that the info collected will truly reply the analysis questions and that the methodology works easily in a real-world distant setting.

Periods ran in UXtweak’s Distant unmoderated mode. Members learn a job immediate, clicked to Google.com/aimode, prompted AI Mode, and spoke their ideas aloud whereas interacting with AI Mode. They got the next instructions: “Suppose aloud and briefly clarify what attracts your consideration as you overview the knowledge. Converse aloud and hover your mouse to point the place you discover the knowledge you might be searching for.” Every participant accomplished seven job varieties designed to cowl various intent classes, together with comparability, transactional, and informational eventualities.

UXtweak recorded full-screen video, cursor paths, scroll occasions, and audio. Periods averaged 20-25 minutes. Incentives have been aggressive. Uncooked recordings, transcripts, and occasion logs have been exported for coding and evaluation.

Three educated coders reviewed every video in parallel. A row was logged for UI components that held consideration for ~5 seconds or longer. Variables captured included:

• Structural: Fields describing the setup, metadata, or construction of the research – not person conduct; embrace knowledge like participant-ID, task-ID, system, question, order of UI components clicked or visited in the course of the job, sort of web site clicked (e.g., social, group, model, platform), area title of the exterior web site visited, and extra.
• Function: Fields describing UI components or interface parts that appeared or have been accessible to the participant. Examples embrace UI aspect sort, together with buying carousels, service provider playing cards, proper panel, hyperlink icons, map embed, native pack, GMB card, service provider packs, and service provider playing cards.
• Engagement: Fields that seize lively person interplay, consideration, or time funding. Consists of studying and a spotlight, chat and query conduct, together with click on and interplay conduct.
• Consequence: Fields representing person outcomes, annotator evaluations, or interpretation of conduct. Annotator feedback, effort score, the place information was discovered.

Coders additionally marked qualitative themes (e.g., “velocity,” “skepticism,” “belief in citations”) to assist RAG-based retrieval. The analysis director spot-checked ~10% of movies to validate consistency.

Annotations have been exported to Python/pandas 2.2. Placeholder codes (‘999=Not Relevant’, ‘998=Not Observable’) have been eliminated, and categorical variables (e.g., appearances, clicks, sentiment) have been normalized. Dwell instances and different time metrics have been trimmed for excessive outliers. After cleansing, ~250 legitimate task-level rows remained.

Our retrieval-augmented technology (RAG) pipeline enabled three levels of research:

• Knowledge readiness (ingestion): We flattened each participant’s seven duties into particular person rows, cleaned coded values, and standardized time, click on, and different metrics. Transcripts have been retained in order that structured knowledge (corresponding to dwell time) may very well be related to what customers truly mentioned. Aim: create a clear, unified dataset that connects conduct with reasoning.
• Relevance filtering (retrieval): We used structured fields and annotations to isolate patterns, corresponding to customers who left AI Mode, clicked a service provider card, or confirmed hesitation. We then searched the transcripts for themes corresponding to belief, comfort, or frustration. Aim: mix conduct and sentiment to disclose actual person intent.
• Interpretation (quant + qual synthesis): For every group, we calculated descriptive stats (dwell, clicks, belief) and paired them with transcript proof. That’s how we surfaced insights like: “external-site duties confirmed greater satisfaction however extra CTA confusion.” Aim: hyperlink what folks did with what they felt inside AI Mode.

This pipeline allowed us to question the dataset hyperspecifically – e.g., “all members who scrolled >50% in AI Mode however expressed mistrust” – and hyperlink quantitative outcomes with qualitative reasoning.

In plain phrases: We are able to pull up simply the best group of members or moments, like “all of the individuals who didn’t belief AIO” or “everybody who scrolled greater than 50%.”

We summarized person conduct utilizing descriptive and inferential statistics throughout 250 legitimate job information. Every metric included the depend, imply, median, commonplace deviation, commonplace error, and 95% confidence interval. Categorical outcomes, corresponding to whether or not members left AI Mode or clicked a service provider card, have been reported as proportions.

Analyses coated greater than 50 structured and behavioral fields – from system sort and dwell time to UI interactions, sentiment. Confidence measures have been derived from a JSON evaluation of person sentiment by way of transcripts of all customers.

Every job was annotated by a educated coder and spot-checked for consistency throughout annotators. Coder-level distributions have been in comparison with affirm steady labeling patterns and inside consistency.

Thirty-seven members accomplished seven duties every, leading to roughly 250 legitimate duties. At that scale, proportions round 50% carry a margin of error of about six share factors, giving the dataset sufficient precision to detect significant directional variations.

Pattern dimension is smaller than our AI Overviews research (37 vs. 69 members) and is supposed to study U.S.-based customers (all members have been dwelling within the U.S.). All queries passed off inside AI Mode, that means we didn’t straight examine AI vs non-AI situations. Suppose-aloud might inflate dwell instances barely. RAG-driven coding is just as sturdy as its annotation inputs, although heavy spot-checks confirmed reliability.

Members gave knowledgeable consent. Recordings have been encrypted and anonymized; no personally figuring out knowledge have been retained. The research conforms to Prolific’s ethics coverage and UXtweak TOS.

Featured Picture: Paulo Bobita/Search Engine Journal