From Scattered Tools to Systems of Support: What Google’s Big AI Announcement Means for Education

From AI-powered feedback to personalized learning pathways, Google’s updates offer a new foundation for deeper learning. Plus an invite to my live strategy session.

Google didn’t just launch new features at this year’s ISTE + ASCD conference, they delivered a system. Lesson planning. Feedback. AI tutoring. Language support. Story generation. Video and podcast creation. All embedded across Gemini, Classroom, Docs, and NotebookLM.

It wasn’t just about adding more. It was about offering something coherent. For the first time, AI is being positioned not as a set of disconnected tools, but as a coordinated layer across the learning experience, planning, reflection, assessment, and communication.

I had a chance to preview the updates before ISTE, and as someone who rarely promotes AI tools, this is one I can’t stop talking about. Here’s why:

Google built the infrastructure. Now the challenge, and the opportunity, is deciding how we’ll use it to design systems for deeper learning.

And if this is the kind of systems work you’ve always cared about then you’ll love this.
So what’s different compared to other technologies that have brought similar promises. First I don’t know that I’ve seen anything like this. Second, 80% of what’s available is free to Google Workspace users (and honestly most people won’t even be ready for everything that is free), and above all this is a set of tools grounded in research, learning science, and enterprise-grade infrastructure in what they call LearnLM.

What is LearnLM?

At the core of this update is a model built specifically for learning: LearnLM. Developed by Google and DeepMind in collaboration with educators and learning scientists, LearnLM is fine-tuned to support how people teach, reflect, and grow. It’s grounded in principles like managing cognitive load, supporting metacognition, and adapting to diverse learning needs.

In a May 2025 evaluation led by 189 educators and reviewed by 206 pedagogy experts, Google conducted a multi-turn head-to-head assessment of major AI models. The results were clear: Gemini 2.5 Pro was preferred in 73.2% of blind comparisons, outperforming Claude 3.7 Sonnet, GPT-4o, and OpenAI o3. It also ranked highest across five core indicators of effective instructional support.

This isn’t just a better interface or faster response time. It’s a significant step toward aligning AI with the deeper goals of education, and embedding those capabilities directly into tools like Classroom, NotebookLM, and Docs. Let’s take a closer look.

Key Highlights from Google’s Latest AI Education Updates

Generate more than text with Gemini. If you have only used Gemini to generate text, you are only using about 30% of the platform. Inside Gemini, especially when you use Canvas you can create websites, infographics, quizzes, docs, audio overviews and more. This is available to students and schools.

P.S. There are a LOT of updates. If this feels confusing, click here to watch my free webinar where I walk you through all of it with examples and context for what this means to you.

Here a student generates an interactive diagram of the anatomy of an animal cell

Notebook LM: The first viral Google tool that put them on the map has a huge upgrade. You can now generate audio and video overviews. With the audio overviews you can even join in the podcast conversation and ask questions. An interactive podcast! Absolutely mind blowing. Again this is available to both schools and students.

Google Vids: This is officially my favorite tool to do workshop reflections with. Simply type in a prompt with your aha moments and watch the magic. You can replace the audio and visual elements and so much more.

Deeper Integration of Gems. If there was just one tool that I believe could change it is making your own custom Gems. Regardless of tool, the skills you learn, the workflows you design, and the opportunities you create for yourself with custom tools trained to act and think like you is unlike anything else. Creating Gems has been around for a while. However, being able to share them, and integrate them into Google Classroom begins to allow everyone to take advantage of the research that says 1:1 support increases outcomes for learners.

Tag Learning Goals in Google Classroom. This is the update I’m most interested in, not because of what it does, but because of what it makes possible.

In the coming months, educators will be able to tag assignments in Google Classroom with specific learning goals, drawing from national standards like NGSS, ISTE, ACT, and state frameworks. But it doesn’t stop there. Through Google’s partnership with 1EdTech and Common Good Learning Tools, any institution can add their own custom frameworks using the CASE Network 2 infrastructure.

This means we’re no longer limited to tracking completion or assigning points.
We can begin to track progress toward actual competencies.

It’s a subtle shift on the surface. But underneath, it opens the door to a different kind of assessment system, one that’s aligned with skill development, not just seat time or letter grades. If you’ve ever asked how AI might support deeper, more personalized learning, this is one of the most important signals to watch.

From Features to Frameworks

LearnLM gives us more than just better tools, it gives us a foundation. For the first time, we have AI capabilities that are grounded in learning science, embedded across Google’s ecosystem, and designed to support real instructional goals.

But even with that progress, the challenge remains:
How do we turn these capabilities into a system of support that feels coherent and sustainable?

Because right now, many schools are stuck in what I call the tool zone:

One platform to generate quizzes.
Another to summarize research.
A third to give feedback on writing.
A fourth to track learning goals.

Each one is useful, but scattered. And in many cases, expensive, duplicative, and disconnected. It is very obvious this substitution type use was not part of their vision. But they added it to bring people in.

❤️ And this is where I am going to hold your hand and say: Yes we need all stakeholder voices at the table. But brining in the voices of people who will simply ask you for things to be done better and faster is not it. And unfortunately that is what not only many educators will ask for, but what people on the periphery of education will ask for as well. If we truly want to solve the challenges that the education profession faces then we need to raise the bar. Because right now it is on the floor.

The irony is not lost on any of us that the same algorithms powering AI that you think are killing critical thinking, are also the same ones that power the viral 30-second teacher hack videos promoting these substitution trap tools.

What Google has quietly done is replace the need for many of these standalone tools, platforms like Magic School, Brisk, and Quizizz with something more integrated:

• A shared infrastructure

• Deep integration across Workspace and Classroom

• Grounded in learning science

• Backed by enterprise-grade privacy and security

This isn’t just about saving time, it’s about moving from fragmentation to alignment, from disconnected tools to systems that truly support people.

But infrastructure alone isn’t enough.

The real transformation begins when we use that foundation to define a vision, one rooted in how we want teaching and learning to evolve so that every individual has the mindset and skills to not only see themselves in the future, but know that they have the agency to influence and shape what comes next.

We need to make it Click, for ourselves, for our teams, and for the systems we’re trying to build.

Make it Click

One of the most valuable shifts we can make right now is moving from scattered action to shared clarity. In their new book Click: How to Make What People Want, Jake Knapp and John Zeratsky offer a framework that helps teams do exactly that. They call it the Foundation Sprint, a way to define the strategic core of your work before the system even exists so that people become excited about what is possible.

I had the chance to get an early copy and meet Jake Knapp and John Zeratsky at their book launch for Click, and it immediately resonated with the kind of work we do in schools. I’ve seen how powerful this approach is for communicating AI initiatives across schools.

Their approach to building clarity before systems gave language to what so many of us have been feeling: that the real challenge with AI isn’t access, it’s alignment.

At the heart of the “Click” process is something they call a Founding Hypothesis: a clear, testable statement that helps your team align around what you're building, who it’s for, and why it matters. It’s not about guessing the future. It’s about creating enough focus to move forward with intention.

Here’s one version of what that might look like for school leaders right now:

If we help school leaders solve the challenge of fragmented AI adoption with a clear, three-part framework and integrated Google tools, they will choose it over expensive edtech platforms and temporary workarounds because our solution is scalable, strategic, and rooted in systems, not just tools.

But behind every hypothesis is a deeper belief.

I believe that if every educator had one AI teammate they could train, reflect with, and rely on, they’d be more creative, more supported, and more able to design learning that puts student agency at the center.

Once It Clicks, It’s Time to ACT

That’s why I created the ACT Framework, a clear, three-step system to help educators and leaders turn powerful tools into meaningful support. It’s designed to guide teams from experimentation to transformation, with a focus on strategy, sustainability, and what learning needs next.

P.S. I’ll be hosting a live strategy session and QA so if you want to join click here.

Let’s break down each phase of the cycle:

Assess: I explore what’s possible before I commit.

Many people say the problem with AI, or any kind of change really, in education is fear. That people just need “step 1.” But the challenge I see is that we often design that first step without steps 2, 3, or 4 in mind. We push people to implement before they’ve had a chance to explore. And that’s how we get stuck in substitution mode, using AI to write a quick quiz, generate a basic lesson plan, or clean up an email. 

We’ve normalized these examples as progress. But we just have to look to the past two decades with other technologies to see that this approach is anything but transformational. 

So with AI, let’s agree, it’s time to stop normalizing mediocrity.

Let’s stop using our own fear as a reason to say others “aren’t ready.”

Let’s raise the bar for what thoughtful, strategic integration actually looks like.

This tension isn’t new. Back in 2001, sociologist Paul Attewell introduced the idea of the digital use divide, the gap between simply accessing technology and using it to do meaningful, higher-order work. He warned that without intentional design, technology would replicate existing inequalities and reinforce shallow habits.

That’s what we’re seeing again with AI.

I’ve been openly critical of many AI tools, platforms like Magic School, Brisk, Quizizz, and others, not because they don’t save time, but because they often narrow our focus. They promise support. And for many, they deliver exactly what’s needed in the moment: a faster way to get through the day. But over time, they can unintentionally limit your imagination, reinforcing old habits instead of helping us design new systems.

That’s the difference between tools that feel helpful and tools that actually lead to change. Sometimes real support doesn’t come from automating a task. It comes from creating the conditions to think differently. Not just giving people what they want, but helping them grow into what they need.

So I can see some of you asking, if you don’t like Magic School, why do you support this being built inside Gemini?

The answer is simple. They designed with the end in mind.

Because Gemini isn’t about one-size-fits-all.

It meets you where you are, but it doesn’t keep you there.

This phase isn’t about proving you’re “good at AI.” It’s about browsing the menu, choosing something that resonates, and letting the tool guide you toward what’s possible. Maybe it’s lesson planning. Maybe it’s generating feedback. Maybe it’s summarizing a resource or helping you think through a tough conversation. The point isn’t to master the tool, it’s to experiment. To notice what feels useful, what sparks a new idea, what makes you pause and rethink. Meaningful innovation starts with empathy, not expectations and real change happens when people are given space to see what’s possible for themselves. When you start to recognize those patterns, what’s working, what’s not, and why you’re no longer just exploring.

You’re ready to shape something more intentional.

And that’s where Customize begins.

Customize: I train my AI team with what works so it scales the practices I trust.

This is the shift from using AI as an assistant to building an AI team. This stage is about technology + designing the culture, norms, and values that will govern how you and your digital teammates work together. According to Microsoft’s 2024 Work Trend Index, phase one of AI adoption is all about productivity gains, asking AI to complete tasks like writing emails or creating rubrics.

Phase two is different. It’s about building custom teammates that understand your work, your goals, and your voice. That’s what happens when you design a custom Gemini chatbot, or train a NotebookLM notebook on your materials. You’re not just saving time, you’re designing a partner. One that carries your expertise forward. One that reflects your instructional or leadership style. One that doesn’t need to be re-taught every time you open a new tab.

When you create a Gem that generates feedback aligned to your rubric, or a NotebookLM workspace that summarizes your department’s goals, you’re doing more than improving efficiency. You’re encoding your best thinking into a system that can support others, and grow with you. This is where AI becomes less of a taskmaster, and more of a teammate.Less of a tool, and more of a design decision.

Transform: We build systems that support people, not just processes.

Every school has that person you go to when you’re stuck. The colleague who always knows the best prompt for reflection. The coach who gives just the right kind of feedback.The teacher whose lesson plans somehow balance structure and creativity every single time. That’s individual talent.

But transformation is about turning that brilliance into shared intelligence. It’s about building AI teammates that don’t just serve one person, but reflect the collective wisdom of a team all in the service of doing what couldn't not have been achieved alone.

• A feedback assistant trained not just by you, but with insights from across your department.

• A NotebookLM space that curates and summarizes resources for everyone, not just those who had time to dig for them.

• A Gem that captures your school’s approach to project-based learning or assessment, so it doesn’t have to be reinvented every year.

When we stop designing tools for individuals and start building systems that learn with us and from us, we stop chasing innovation in pockets. We scale what works. We make it accessible, repeatable, and resilient.

This is what today’s world is asking of us: To explore with curiosity. To design with intention. To build systems that let humans and AI work together to scale what matters most. Not just doing more, but making possible what we couldn’t imagine alone.

The ACT framework is a mindset shift, but it’s also a practical roadmap.

And now, with Google’s latest updates, we have the tools to bring it to life.

Live Strategy Session + QA: How to Build Your AI System with Google Gemini

This year I’m going back to school with you. I’m teaching at UC Irvine in the “Technology and Media Literacy” class to Masters and Credential students and I am SO excited. I’ve also been integrating this approach with our K-12 district partners.

In this free live strategy session I’m going to show you:

• Exactly how I am thinking about building my AI team

• How I am thinking about helping these soon to be teachers build theirs too

• How our partner districts are implementing the ACT framework

• And I’m going to answer any questions you have about building yours.

• I’ll show you how to apply the ACT framework using Gemini, NotebookLM, and Google Classroom, so you can build systems that scale your team’s best thinking, not just speed up your to-do list.

Click here to register for the free live strategy session.

This is the year it can be different.

As Seth Godin says, “We don’t have to be the victims of a system that has outlived its usefulness.”

Final Thought

The future of education won’t be built by tools. It’ll be built by teams that are both human and AI. And if we want AI to be a real teammate, not just a tool, we need more than access, we need strategy, systems, and leadership.

We can build something better.
Let’s do it together.