Like It Or Not, Google Is Slowly Stair-Stepping Us Into AI Mode
It’s hard to believe but soon we will be coming up on the 2nd anniversary of AI Overviews. As you can imagined a lot has changed since then in the search results. Google has been aggressively pushing out AI-driven features and there’s no sign of that slowing down.
After multiple conversations with my college Greg Boone at Walk West, he made a point that’s resonated with me ever since. Google has been “stair-stepping” it’s users into AI Mode. He believes Google has no choice but to adapt into a full-on LLM. I 100% agree with his perspective.
However, Google’s problem is they couldn’t simply change the search results overnight. They would risk untold amount of ad revenue, massive user churn etc. Instead, they have to slowly “educate” the users on how to use and interact with an LLM. Overtime, as users get more comfortable with the concept, Google can begin to roll out the changes they wish they could’ve years ago.
In order to show everyone this, I’ve constructed a timeline of just how Google is doing this. Worth noting that I couldn’t put this together without the relentless reporting from Barry Schwartz on Search Engine Roundtable. His documentation of each event makes it possible to string together.
So let’s go over the timeline.
The AI Mode Timeline:
Nov 30, 2022: OpenAI Launches ChatGPT
In the event that kicks this whole series off, OpenAI launches ChatGPT in November 2022. It’s released as a “Research Preview” in what will be known as one of the most pivotal marketing launches of all time. Now your everyday user has access to an extremely powerful LLM.
Dec 21, 2022: Google Issues A Code Red
Less than a month after the release, Google management issues an internal “code red” to its employees. They quickly realized the threat that an LLM could bring to their search business unlike any other technology they’ve seen before. Their team is quickly made to “to switch gears to assist in the development and launch of AI prototypes and products”. This is when the aggressive changes to AI results begins.
May 10, 2023: Google Releases “Search Generative Experience”
By the time Google I/O 2023 rolls around, Google has released the Search Generative Experience prototype (SGE). SGE is a “half way” approach. It blends the insights of an LLM but throws that on top of the traditional search experience. It’s a way for users to see how an LLM could work while letting the “10 blue links” exist.
Despite it’s “rougher” design, it’s actually a pretty decent proxy for how “AI Overviews” looks today.
May 15, 2024: Google Launches AI Overviews
One year later at the Google I/O 2024, Google fully rolls out SGE as AI Overviews to the US market. Fundamentally, it’s not that different than SGE. However, now it’s shown by default in the search experience.
Google frames AI Overviews as being able to “help with more complex questions.”. Basically teaching users that they can search for longer queries and they’ll now get an answer, similar to an LLM.
Oct 16, 2024: AI Features In Google Shopping
Google announces that they are making massive overhauls to Google Shopping, including many new AI features. While it’s not a directly connection, we can basically see they’re using AI Overviews technology as part of the commerce experience. Now they’ve expanding LLM functionality to the Shopping ecosystem.
Nov 7, 2024: AI Overviews In People Also Ask
Now Google begins to expand AI Overviews into more subtle places in the search result. They integrate AI Overviews, into People Also Ask (PAAs), which appear in prominent placements in every search result. Now expanding those options gives users another answer from AI, not from sites.
Today’s PAA’s still represent a mix of standard sites + AI results.
Mar 19, 2025: AI Overviews Expand To Middle Of Results
Now Google start presenting the AI answer across different places in the search results. Now instead of at the top results, AI Overviews can start appearing in the middle (and eventually bottom) of results. This likely was an easier way for them to eventually start expanding the number of AI Overviews they displayed.
Apr 8, 2025: Google Increases AI Overviews
Now instead of just showing them for “complex queries”, Google start rolling out AI Overviews more aggressively. SEO tools such as Ziptie.dev report that over 20% of queries show an AI Overview.
They need more users to interact with AI features in order for them to get familiar with. Expansion is the most logical way. This eventually leads to The Great Decoupling as more sites increased impressions but static/dropping clicks.
Jun 13, 2025: Google Rolls Out AI Mode
At Google I/O 2026, Google now makes another huge push. This time they announce “AI Mode” will be rolling out to all US users. AI Mode is a full-on LLM with no “10 blue links” to be seen. In the main search results they’re a prominent “AI Mode” tab that users can select. This is another subtle but powerful step towards full integration.
This is where Google wants people to end up but they can’t fully integrate it to the search results yet. Just like they did with AI Overviews, they “set it off toe the side” of the main search experience and let users interact with it on their own accord.
Jul 21, 2025: AI Mode Button Comes To The Home Page
Google makes another subtle but strong change. They add an ” AI Mode” button on the home page of Google search. A small change but it makes AI Mode feel more integrated and normalized into Google search. If it’s part of the home page, it must be one of the key features of the product.
Aug 21, 2025: AI Mode Rolls Out To 180+ Countries
Like they did for AI Overviews, they need more scale for AI Mode. As a result, they release AI Mode in 180+ countries to give global access to the product. Now worldwide users can have access to their LLM and learn how to use one.
Sep 1, 2025: AI Overviews In Knowledge Panels
While AI Mode is happening, Google keeps expanding where users can see AI Overviews in the search results. Now on top of the PAAs, they add in Google Knowledge Panels. Another subtle change that shows Google giving an “AI-driven” answer instead of something they obtained from their knowledge base.
Dec 2, 2025: Google Integrates AI Mode Into AI Overviews
This is one of the biggest changes but once again very subtle. Google now start integrating the two products together. They add an “AI Mode” button to the bottom of AI Overviews.
Instead of having AI Mode silo’d off from the rest of the results, they integrate it to the AI Overviews directly. Now users can “flow” from one feature to the other seamlessly.
Dec 3, 2025: Image/File Uploads Take Users To AI Mode
One day later, another integration of AI Mode happens. Now Google adds features to their search bar for users to upload either images or files. Performing either action pulls the user directly into the AI Mode experience.
Now they train users that search can be a multi-model experience. Instead of text, they can interact with search via images, PDF etc. However, this teaches users that an LLM is the medium that answer takes place, not traditional search.
Jan, 28, 2026: Follow Up Questions In AI Overviews Integrate Into AI Mode
Google follow up with yet another integration of AI Overviews + AI Mode. On mobile devices, after expanding an AI Overview, users can ask even more questions. However, those questions roll them directly into the AI Mode experience.
So now Google has engineered a marketing funnel into AI Mode. Now a user start a standard Google search. But then they get an AI Overview as the answer. They’re also shown that they can ask follow up questions to this search, like an LLM. Asking these questions rolls them directly into the LLM-experience.
What Happens At Google I/O 2026?
So looking at the evidence, I’m hard pressed not to imagine an announcement where Google rolls out AI Mode. For the past two years, they’ve been training users with AI Overviews. Teaching them that they can get answers to extremely long search queries. They’ve slowly dripped that functionality into different aspects of the search results (PAAs, Knowledge Panels, Shopping). They aggressively started showing more AI Overviews during that two years time.
Now they’ve been integration AI Mode into the AI Overviews experience. Continually pushing more users towards this ecosystem whether it’s through UX changes (AI Mode tab, home page button) or direct integration (questions, files/images button).
After announcing SGE back in 2023, they only waited one year to roll out AI Overviews to the default search experience. Well come May 2026, AI Mode will have rolled out for about a year.
Sounding The Alarm Bells
So I’m here to give the bad news. Whether you like it or not, AI Mode is coming. Very likely, it’s this year. This means that traffic will DRAMATICALLY drop. SEOs will basically be forced to drop reporting on organic traffic as mentions + citations will be the only meaningful metric that AI Mode will produce.
Those saying that “organic is still 80% of our traffic” might see how overnight that drops off a cliff. Google has the power to change into LLM overnight and they will very likely decide to pull the trigger.
How To Prepare For AI Mode
Is all hope lost? No, not at all. In the immortal words of my co-founder Jason Melman, customers are still looking for products and services to buy through different platforms. While the functionality and metrics of those platforms is changing, the behavior isn’t.
So what can you start doing?
- Play around and test AI Mode. Familiarize yourself with the interface, how it works and how it cites sources.
- Read Mike King’s “How AI Mode Works“. It’s one of the most comprehensive tear down of the technology.
- Start tracking key prompts in an AI Tracker. At Nectiv we have our own internal tracker but there’s also Profound, Athena, Peec and others
- Research how fan-out queries work for Gemini 3. We’ve done a lot of research around this and can even pull some queries for you.
- Be honest about your strategies. If you need to integrate third parties (Reddit, YouTube, Affiliate), don’t be afraid to do so and be honest about it.
- Look to understand how your customers use conversation, not search. We wrote an article about how you can mine that in Search Console.