Everything We Know About Google’s AI Mode

On March 5th, 2025 Google launched a new “Labs” experiment that might signal what a near-future, fully AI version, of Google will look like called “AI Mode”. Shocking to few, this new mode of Google greatly resembles the AI experience pioneered by OpenAI’s groundbreaking ChatGPT in November of 2022 and leaves little to no room for downstream click traffic to websites, especially those who originally published the content being plundered by Google’s AI systems to create this new mode.

This article is a collection of everything we know so far about AI Mode and it includes a few educated guesses as well.

Everything we know about Google’s AI Mode:

  1. This is being framed as an experiment and it may never reach the masses.
  2. AI Mode is powered by a custom Gemini 2.0 model and Google claims it has advanced reasoning capabilities.
  3. Google’s announcement includes features that do not currently exist and may not exist in the future stating “Note: Results are for illustrative purposes and include future-looking features. This example reflects information from a prior date, and may vary.” in their original post.
  4. Users have to opt in to AI Mode, which will mean low adoption until if/when Google forces it on users.
  5. Like the highly invasive AI Overviews, AI Mode does not directly include links from the results, but instead uses a link icon (which no internet users know the meaning of) that opens a card full of various links which a user then has to peruse and click to go anywhere. (note: in Andy’s video embedded below it does look like there are blue links, so maybe in some cases they are included?)
  6. AI Mode appears as a new tab on a Google search result, once toggled on users will see what is similar to an AI Overview. Once opened the AI Mode component includes a colorful button that says “Dive Deeper in AI Mode” and when clicked opens up a more ChatGPT like experience.
  7. Citations in AI Mode are extremely small and easy to miss, this is on par with AI search engine Perplexity, though Perplexity’s solution appears to do a much better job of citing the actual source used for a segment of an AI response.
  8. While currently links in the results are scarce, demos from Google and some screengrabs from early users show content that Google sources being displayed in card formats like many other experimental parts of Google (i.e. Shopping) there is little data on how users interact with cards versus links.
  9. AI Mode appears to have access to Google’s various sets of data including local business data from Google Business and news from Google News.
  10. There currently is no analog to the “3-pack” in normal Google SERPs for local SEO queries, instead AI Mode just spews out location data matching the request, likely from the Google Business database.
  11. AI Mode click data is NOT included in Google Search Console reports for now.
  12. AI Mode buttons are already appearing in beta versions of the Google app.
  13. Some users are saying Google’s AI Mode is more accurate than Perplexity.
  14. Some users are saying Google’s AI Mode is faster than Perplexity, ChatGPT, and other competitor products.
  15. SEOs are torn on what happens to their industry if Google replaces the SERPs we’ve all known for 27-years with AI Mode.
  16. Google VP who announced AI Mode, Robby Stein, stated on X that the team wants AI Mode to continue driving clicks to the web and that he believes it will create new opportunities for websites in the future.

About the Layout
Google’s AI Mode looks a lot like a really big version of AI Overviews with the option to chat with Google at the bottom. This effectively combines the AI Overviews look and feel with a bit of ChatGPT or other AI Chatbot functionality.

Inside we see a stark white page filled with text instead of blue links with extremely tiny link icons. Once clicked a link icon opens up a card showing links to content that Google used to show the information to a user off to the right hand side on a gray background with the text and url all looking dull and unclickable / untappable.

The entire layout beckons you do to do nothing, there is no feel like clicking/tapping a link or icon will do anything for you, the prompt at the bottom hacking human psychology to try and ask something again instead of visiting one of the several sources who spent their own time and money collecting data and writing the article or making the video.

It is clear here in this image of the first version of AI Mode who wins if AI Mode takes over Google’s dominant 90%+ of global search share, there isn’t even a remote guess. Google wins, publishers (especially the independent and small ones already reeling) lose big time. Only those that Google has to send traffic to such as brick and mortar companies, ecommerce, or specialty places have a chance at winning the traffic on the web that will remain once Google fully incorporates all of their platforms and content/data from websites.

More on Google’s AI Mode

AI Mode appears as a new tab on a Google search result, once toggled on users will see what is similar to an AI Overview. Once opened the AI Mode component includes a colorful button that says “Dive Deeper in AI Mode” and when clicked opens up a more ChatGPT like experience.

AI Mode click data is NOT included in Google Search Console reports for now.

There currently is no analog to the “3-pack” in normal Google SERPs for local SEO queries, instead AI Mode just spews out location data matching the request, likely from the Google Business database.

SEOs are torn on what happens to their industry if Google replaces the SERPs we’ve all known for 27-years with AI Mode.

Some users are saying Google’s AI Mode is more accurate than Perplexity.

Some users are saying Google’s AI Mode is faster than Perplexity, ChatGPT, and other competitor products.

Google VP who announced AI Mode, Robby Stein, stated on X that the team wants AI Mode to continue driving clicks to the web and that he believes it will create new opportunities for websites in the future.

Speculations
These are things I think MIGHT happen with AI Mode or just Google’s AI in general.

1. Say hello to the Ads
As Google dispatches with competition and/or purchases them, ads will start creeping into and above AI Mode. It’s a stroke of genius if Google can keep the regular SERPs with AI inserted into them which will have its own ads inside of it. Exactly what Mr. Raghavan wants, every search now has to trigger at least 2 ad impressions and advertisers have no choice but to pay for those – likely no targeting options to focus on one or the other either. Not exactly a ground breaking prediction considering Google’s past with ads, I know, but one to be aware of – especially for users who might assume a clean, ad free, AI Mode interface will always exist (hint it won’t).

2. Perplexity had a good run
Perplexity is likely the first of the native GenAI discovery systems to fall. Building revenue models signals they are most likely running out of direct investor support and need to generate their own revenue in order to keep closing rounds. If Google’s AI Mode steals a good portion of their users away my best guess is they close or sell to Google seeing no other option to keep going.

3. Publisher Collapse
Most publishers are barely making it right now. As AI Overviews expand and as Google’s AI Mode increases in usage, the number of outbound clicks to read a small portion of a news article will also collapse. Those Google has a contract with or those with a solid enough base of subscribers / other revenue streams will survive. Those that rely on advertising impressions from organic are going to have an increasingly rough time and many smaller ones will close shop for good. If Google starts linking better inside of the content or does other things to encourage clicks to downstream content this might change, but at this point in the GenAI / ChatAI game it doesn’t seem like a good move for Google to make since the hype typically comes from internet users who hate ads and really don’t like publisher websites that tend to have a lot of ads.

4. More Color and Cards
Google’s announcement video which appears to be getting shared the most out of all related media so far, shows a much more colorful AI Mode than what users are currently experiencing. I suspect we’ll see colorful inserts from other Google owned properties such as YouTube, Hotels, Flights, Business / Maps, and Shopping / Merchant Center soon. There’s even a chance we see updates to how Google allows users to click-through to content or how Google presents content itself. For example if an AI query is about a recent trending event Google might include article bubbles for a user instead of a the much less click-worthy gray link icon.

I could also see a version of this where social media content gets inserted if Google’s AI thinks the user wants to engage with other humans.

IF AI Mode expands in this way I believe it gives Google an edge over competition and aligns with early statements from VP Robby Stein (which go against Google’s earliest statements on GenAI experiences btw).

5. Highly Qualified Customers On The Way
If Google wants to keep the web healthy and thriving and If Google determines that users need to go downstream at some point, and consumers use AI Mode to ask a lot of questions, then we might see something like AI Mode become an organic lead filter system that qualifies leads before they even hit your site. This would, of course, be monetized by Google in some way eventually. It would also require some way of Google knowing what your perfect customer looks like, that could be an expansion of ads.txt, a new file, data inputs somewhere on Google, or just via really specific content on your site (i.e. “Our widgets are perfect for moms over the age of 40 who like to do Yoga and work an outdoors job”). For upset businesses, this will probably be how Google sells to them to not be as unhappy with the move. Traffic on the open web will dwindle but in such a scenario value might stay flat or even increase. Google will convince us that its better for us to write content that Google then steals via GenAI to prove our expertise and filter for highly qualified leads than for that traffic to come to our websites and give us a chance to market to them ourselves. This also places Google in an even more powerful position (which seems almost insane) over web-based companies.

6. Google Likely WANTS AI Mode to Replace / Shove Down SERPs
Since the earliest days of Google when Aaron Wall started showing a slow but steady pattern of Google expanding the screen real estate their ads took up, Google has pushed down organic content in hope of keeping eyeballs and clicks for advertisers. AI Mode is no different and I wouldn’t be surprised to see it taking up more space over time or even to randomly appear in normal Google SERPs at various ranking positions. Like most platforms Google wants to keep users on Google until they absolutely have to leave, it makes them far more money that way and I see AI Mode / Overviews playing a role here eventually if consumers adopt it.

7. Google’s Algorithm is Still Important
Most LLM-AI systems are really similar with a few changes here and there. Elon Musk’s Grok is more laid back and can curse, OpenAI’s ChatGPT tries to be as proper as possible, etc… A system like this needs something unique – just like the early days of meta search engines – or they’ll run the risk of being just too similar to everyone else and will never stand out enough. For meta search engines the answer was Backrub / Google’s algorithm “PageRank” which has grown over the past few decades and at least for now that current system is likely a big advantage for Google – if they can find ways to reward content quality and originality not just the most popular websites.

8. Google is Hungry
With AI Mode Google is showing us that they are hungry, not for the same content everyone else has, but for unique content. The more likely Google is to answer a question with your content and the user be satisfied more than other AI systems, the more likely you win downstream traffic / actions. This isn’t a reality yet, unfortunately, but likely is on Google’s radar as they look for ways to edge out rivals and push ahead like they did back in 1998. Real humans, creating really unique, and creative but accurate and useful content would be that advantage. They stink at this right now and if AI Mode signals anything it is that Google believes they are ready to pull ahead in this aspect beating out the more stale content stores of their competitors.

My Predictions About “AI Mode” / AI Overtaking Search
I have been predicting “AI Mode” or whatever it ultimately becomes called since around 2012 when I worked at an SEO agency and realized how close the tech industry was to figuring out human speech and the implications that came with this pending breakthrough. Since that time I have written about the topic and when given a chance even presented it about it SEO conferences, though to be quite frank most – including my favorite Pubcon – never accepted my pitches to talk about this potential now seemingly imminent future. Some notable moments from that journey.

In 2012 I spoke about this prediction to executives at the marketing agency I was at, one was annoyed with my constantly trying to figure out the future of things and mocked me, then demanded that I write it down joking about how he wanted to read it later. I never would have started this website or my own agency without that mocking insult. Eventually I decided I didn’t care if I was right or wrong and wrote about parts of this future in a post predicting the emergence of what we now call “smart speakers”, my prediction was 2-years before Amazon announced Alexa and 4-years before Google got around to launching Google assistant / home – though my blog post did not come until early 2014. As far as I am aware the only people who predicted smart speakers existing at the time were myself and Jeff Bezos.

In 2016 I gave a talk at a show in Oklahoma City where I discussed the peak of Google likely being that year and a looming downfall. That the quality of SERPs would be detrimentally impacted, more ads would pollute Google’s products, and even PPC advertisers would start to lose keyword data all as part of this build up. I discussed concepts like the predictive brain and models about brain / consciousness storage and that this was where Google was pushing their tech. Collecting data sources to develop deeper understandings of human speech, knowledge, and cognition in order to eventually create what we now call AGI. In that talk I spoke about Google wanting to do all of the searching for a consumer and while I wasn’t quite sure how they would do I mentioned using in-home sensors as one possibility and another being they would have software perform actions on a human users behalf with just a simple request. Duplex was announced just 2-years later, another stepping stone in Google’s path to being a Task Completion Engine, though there is not yet any in-home sensors by Google capable of this the underlying tech already exists (i.e. doorbell cameras and computer vision AI tech) it wouldn’t take much for some tech company to start combining off-the-shelf tech and building such a sensor.

In 2017 I talked to a crowd in Houston about a “Web without websites” and the looming “War of the walled gardens” which would eliminate traffic on the web and damn value up at major tech corporations. That crowd was shocked when a Google representative doing Q&A from the stage after his closing keynote admitted that such a concept had been considered as viable by Google but cautioned against overreacting (I am still unable to find that footage if anyone out there has it, let me know please). In my talk I urged the attendees to stop building websites and to instead start focusing on building a “Batman” a term I used along the comic book theme of the talk to represent the perfect, the ultimate version of something (i.e. Batman is the perfect private detective / non-super powered hero).

In 2018 I wrote a blog post detailing what I believed Google’s final form would look like, a moment in time before humanity had instant access to everything and a possibly far more open system replaced Google completely. I had talked about this privately to clients for years, helping position them to be ready for this future in ways that they hadn’t considered and wouldn’t have considered otherwise. For example ensuring that clients worked to be named “the best” by any local media in any competition since these future decision making, task completing system might use that information, gathering the best reviews possible (also great for SEO of course), ensuring their website allowed users to take some kind of action – a no brainer for most but a lot of ‘brochure’ sites still exist as information only portals. I went as far as to tell publisher clients to start selling something, anything, or to immediately build a subscriber system because ad revenue was about to collapse from all sources. Think about it, a system that knows you have an anniversary coming up, books your dinner reservation, reminds you to dress nice, sends a car to pick up you and your spouse, delivers you to the restaurant, pays the bill for your, and arrives to pick you back up and take you home – all done for you automatically or with a simple verbal command.

In 2019 after yet another pitch to discuss the future of consumer discovery systems (i.e. search) was rejected by conference organizer I decided to take a sabbatical from speaking at conferences altogether. I gave the talk I was assigned, but had zero energy or enthusiasm for it and decided to spend the entire year of 2020 at home and with family and friends. That same year I predicted the USA was wide open for a pandemic to strike, but assumed it wouldn’t hit for another 5 or so years, and when a biotech vaccine maker failed a phase III clinical trial I bought their stock at a rock bottom price, Novavax’s price would rise over 12,600% over the next year and a half when I sold at $325 a share.

In 2020 sitting out SEO conferences and seeing my pandemic prediction come true, I decided to believe in myself once again and I dug deeply into AI research papers and determined that we were closer than most thought for “search” as we knew it being prepared for a fundamental shift. As the pandemic waned I blasted out pitches with renewed energy this time avoiding any of the stale, standard, SEO talks like “how to build links” etc… for fallbacks. I was told by several conference organizers that priority went to those who did virtual presentations during the hard times (I declined to do those), that there weren’t enough speaking slots, and ultimately as I persisted I was told to go work for a major brand or write a book. Instead, I decided to stop caring about speaking at conferences and started only giving my future insights to clients. Over the past few years our internal predictions about what happens over the coming year have been nearly perfect.

Since 2020 our team has been building the future I believe is coming instead of talking about it, in hopes of getting in front of it for our clients and ensuring victory no matter the conditions of the field of battle. None of those predictions are public and many have been missed or are being held secret by others – a dark forest filling in the gaps of spaces between the walled gardens as AI overthrows the open web and sends it into disarray with a misalignment of incentives. The maker – taker problem, as it is known in open source communities, put on steroids. Infinitesimally few benefiting from the efforts of the vast masses. A new web gestating, waiting for the right moment to spring into existence.

I am now sharing some of those insights via a Substack Newsletter if you want to subscribe:

Resources / References
1. Robby Stein – VP of Product, Google Search “Expanding AI Overviews and introducing AI Mode”: https://blog.google/products/search/ai-mode-search/

2. Barry Schwartz – Search Engine Roundtable “Google AI Mode – It’s Here – Here Is How To Access It”: https://www.seroundtable.com/google-ai-mode-39002.html

3. CNBC Television “Google unveils ‘AI mode’ in search”: https://www.youtube.com/watch?v=E5zC9cGx2sI

Joe Youngblood

view all posts

Joe Youngblood is a top Dallas SEO, Digital Marketer, and Marketing Theorist. When he's not working with clients or writing about marketing he spends time supporting local non-profits and taking his dogs to various parks.

0COMMENTS Join the Conversation →