Google’s keyword team has published its summary of the biggest AI announcements since April 2026. Cloud Next ’26 introduced the Gemini Enterprise Agent Platform and Google’s eighth-generation TPUs, designed for agent workloads. Google also released Gemma 4, described as byte-for-byte the most powerful open source model available, as well as Deep Research Max for improved data integration and a new Colab coding instructor.
The infrastructure numbers are real. Models now process more than 16 billion tokens per minute with direct API use, up from 10 billion last quarter, with nearly 75% of Google Cloud customers using AI products. Developers have downloaded Gemma more than 500 million times, according to Google’s April 2026 AI update.
Trend: A New Type of Search User Is Emerging
In a recent piece based on an episode of Search Off the Record with Martin Splitt and Nikola Todorovic of Google, Google revealed that there is a new wave of people doing things with Search that are significantly different than before, and that this trend is upward. Splitt noted that AI in search has always been behind the scenes, helping organic results. It has recently been ported forward, where it now helps users with increasingly multimodal queries.
That distinction is very important. These are not power users who get the new feature. They are regular users who develop new search behaviors, and those behaviors converge. New users create long conversational questions, and while AI has democratized access to information, at the same time it has made insights based on experience more valuable – something AI cannot easily replicate.
Supporting data reinforces the scale of this change. BrightEdge research found that submissions to AI Overviews grew by 58% in the 12 months to February 2026, B2B technology queries that triggered AI results jumped from 36% to 82% and education queries from 18% to 83%. Those are not incremental changes. Those that are built.
What Bill Ziff Has to Teach Us
Early in my career, I worked for William B. Ziff Jr., the publisher who built the Ziff-Davis empire into one of the most powerful media companies in American technology. He had a saying that I have never forgotten: “People pay too much attention to events and not enough to trends.”
He built his business on that difference. While competitors were chasing the shrinking audience of general interest magazines, Bill Ziff identified a major, structural shift toward specialized technical knowledge and created PC Magazine and a dozen other titles that shaped the way an entire generation learned about computers. He was not responding to news. He was tracking where the audience was going.
That framework is exactly what SEO experts, content marketers, and entrepreneurs need right now.
The Google Keyword Blog has a specific purpose. It keeps experts informed, shows where engineering resources flow, and occasionally contains really useful tactical information. Read it. But don’t confuse it with strategy.
The Gemini Enterprise Agent Platform is an event. Developers who downloaded Gemma 500 million times the event. A new generation of searchers learning to treat Search as a conversational research tool – and expecting answers instead of links – is a trend.
Bill Ziff’s counterintuitive insight was that while events are spectacular, trends indicate where money, audience, and influence actually move over time. The structural change taking place in search right now is behavioral, not infrastructural. Google can send the eighth-generation TPUs and the context window of a million tokens, but the important thing about the content strategy is that users switch to research topics, when the website link does not provide clear answers, they gradually become beggars.
What This Means for Your Strategy
When a new wave of users discovers that search can handle complex queries, and that discovery is a rising trend, three things follow for operators.
First, content that serves those users well – specific, experience-based, specific, designed for machine understanding – will be more important than content that is only optimized for general level signals. AI makes basic information content marketable. He can’t replicate an idea gained through real experience.
Second, the audience itself is changing. Users who ask complex chat questions behave differently from users who type in three keywords. They have high expectations, long timelines, and different conversion patterns. Understanding what changes in your stats is more important than reading about it in a product review.
Third, key metrics are changing. Citation frequency in AI generated responses is becoming more important as was keyword placement in 2015. That’s not a guess – it’s a measurable, trackable signal.
Google’s April announcements tell you what the infrastructure looks like. The new wave of AI users tells you where the audience is going. Follow the audience.
Additional resources:
Featured image: SvetaZi/Shutterstock
