From Rocketship to Flourish, John Danner explores how artificial intelligence is reshaping instruction—and what that means for teachers, students, and systems.
INTERVIEW | by Victor Rivero
John Danner has a habit of showing up early to the future—and building it.
From co-founding the world’s first internet ad server to pioneering blended learning in K–12 schools, his career has consistently traced the edge of transformation in both technology and education. Now, he’s focused on what he believes is the next—and most profound—shift yet: the rise of AI-first schools, where much of what defines traditional teaching is fundamentally reimagined.
Danner’s path to this moment has been anything but linear. After leading NetGravity through its IPO and $750 million acquisition, he stepped away from Silicon Valley to become a public school teacher—an experience that reshaped his understanding of both learning and systems. That pivot led to Rocketship Education, the school network he co-founded in 2006, which became a national model for blended learning and helped demonstrate how technology could personalize instruction at scale, particularly for underserved students.
But if Rocketship helped define the last era of edtech, Danner’s current work is aimed squarely at the next one.
Across ventures like Zeal Learning and his investment firm, Dunce Capital, he has focused on scaling access to high-quality, individualized learning—backing companies such as Outschool, Maven, and SchoolAI along the way. More recently, he has turned his attention to artificial intelligence as the foundation—not just a tool—of schooling itself.
Through ProjectRead.ai and SparkSpace.ai, Danner is exploring how AI can deliver continuous, real-time feedback in reading and writing, unlocking levels of personalization long considered impossible. And with Flourish Schools, the microschool network he co-founded in 2025, he is putting those ideas into practice: designing environments where core academic instruction is increasingly automated, freeing educators to focus on what drew many of them to the profession in the first place—human connection.
At Flourish, that shift is intentional. Danner emphasizes the development of “SuperPowers”—curiosity, agency, persistence—arguing that in an AI-driven world, these human capabilities will matter more than ever. It’s a model that doesn’t just integrate technology into classrooms, but redefines what classrooms are for.
Alongside his work as a founder and investor, Danner teaches at Stanford University, helping the next generation of entrepreneurs navigate the complexities of building in education. Across each chapter of his career, one theme remains constant: a willingness to challenge assumptions—and to rethink, from first principles, how learning should work in a rapidly changing world.
You’ve spoken about the mission to accelerate school systems from a “Not Interested” stage to an “AI-First” reality. For the district leaders and innovators reading EdTech Digest, what does the actual blueprinting of an AI-first school look like in practice, and how does it fundamentally differ from the 1-to-1 device initiatives we’ve seen over the last decade?
As I mentioned in the piece, I think there is a continuum from not interested to strategic use of AI. I don’t expect a school system to act like Flourish soon because we don’t have hundreds or thousands of people to reorganize. These big changes give a huge advantage to starting something new. However, we hope to be an example of the goal where systems can get, and we hope it takes five years, not twenty.
The most reductionist way to say what I think the change in the classroom looks like is: Everything you are doing in the classroom today should be and will be automated. I say “should be” because almost no one goes into teaching to become a content master. They go into it because they love working with kids. And once we get the instructional delivery piece of the job automated, we can get back to relationships. We are about to enter a golden age for teaching, where positive relationships with children are the dominant goal for teachers, something we really haven’t seen since the standards era began.
‘We are about to enter a golden age for teaching, where positive relationships with children are the dominant goal for teachers, something we really haven’t seen since the standards era began.’
A few points on how you automate everything. First, an AI Coach like we have is capable of helping students in about 80% of K–8 standards by having them work through problems and giving them feedback on the fly. While a human teacher likely has more expertise, this tight feedback loop characteristic of 1-to-1 tutoring is superior. This is especially true in areas like writing, where both mechanics and composition are learned by trying, making mistakes, getting feedback, making changes, and getting more feedback. An AI Coach can give 50x more feedback than a human teacher working with a whole class of students.
Areas like planning, grading, providing motivation, and answering clarifying questions for students can all be automated. We had the internet go out one day, and our lead teacher said it was the longest day of her life, because we had taught students curiosity and the desire for feedback as they worked, and suddenly she was again at the center of instruction with every student’s hand raised.
This is fundamentally different from anything that happened with edtech and 1:1 devices, because those tools were at best a supplement to instruction, whereas an AI Coach is a replacement. We shouldn’t be sad about that. If you ask our lead teacher, the ability to work individually and in small groups of students any time she sees the need is one of her great joys after 18 years of traditional classroom teaching.
At Flourish Schools, you’ve described a shift where technology moves from a supplemental “Tier 2 or 3” tool to a primary “Tier 1” instructional driver for foundational skills. In this model, how do we redefine the teacher’s role so it is viewed not as automation replacing staff, but as technology enabling the “luxury” of meaningful, one-on-one human connection?
I think this is the fundamental question, and the one causing a lot of fear among teachers. I get it, but teacher job loss is not what’s going to happen. Instead, we are going to redefine a teacher’s role to connect with students. I think “luxurious” is the right word for modern teaching after the teacher-instruction era squeezed relationships out of the classroom. How many times do teachers get into a great conversation with their classroom but then have to “move on” because the pacing guide doesn’t have room for depth? Flourish is the opposite. Because instruction is now AI-driven and 1-to-1, we have lots of time for depth. Our afternoons are all project-based, and one of the main things we encourage is for students to go deep in the places they have a passion. That means that teachers get to share those passions and have meaningful conversations with students about them.
‘…we encourage is for students to go deep in the places they have a passion.’
Your company, ProjectRead.ai, uses generative AI to create infinite decodable stories tailored to a student’s specific phonics gaps. Given the national urgency surrounding the Science of Reading—or recent attention to the importance of literacy—how can this type of real-time, individualized content generation bridge literacy gaps in a way that traditional, static curricula simply cannot?
Just to give the caveat, I was a co-founder of Project Read, but my co-founder Viv Ramakrishnan is the CEO now and it’s his company to build. That said, we use Project Read’s AI Reading Tutor every day at Flourish. It is by far the best way for our struggling readers to learn to decode because it implements the Science of Reading approach incredibly well, has infinite patience, and a perfect memory of every student’s reading strengths and weaknesses.
At Rocketship, we had amazing teachers, among the best in the country. I would say that half were great with kids, but only half. That was because we had to select for content and pedagogic skills like reading instruction. At Flourish, because Tier 1 reading instruction is automated with Project Read, we can select for teachers with great human skills—their ability to connect with students and create great classroom culture. That’s such a pleasure for me, because classrooms feel great when they are run by teachers whose SuperPowers are in the human rather than pedagogical area.
Edtech has long struggled to measure soft skills with rigor. You’ve branded these at Flourish as “SuperPowers”—qualities like curiosity, agency, and persistence. From a leadership perspective, how do you ensure these AI-driven assessments provide the high-level data that administrators require without losing the joyful, human element of the student experience?
I talk a lot about this here. A few things I would say. First, as educators and parents, we all know that human skills have always been just as important as hard skills. The student who knows the most chemistry is likely at a disadvantage to the most curious student or the one with the most agency. We call the human skills SuperPowers because we want to make it clear that we think they are more important than hard skills—standards. In the age of AI that we are entering now, my opinion is that human skills will be the main differentiator as people are building their lives and careers. Whatever makes us more human is the path to success. Flourish is really an attempt to balance standards and human skills for this reason.
‘…Whatever makes us more human is the path to success. Flourish is really an attempt to balance standards and human skills…’
Here’s the thing. Because human skills were not measurable before AI, “serious” school leaders really couldn’t build the improvement loop so critical to schools growing their students. Now we can measure human skills. Our AI rates students on each of our 24 SuperPowers and is approaching a point where the correlation with teacher ratings is very high. Soon, our teachers won’t need to second-guess the AI rating, because the reasoning behind it is aligned.
That’s a big change. I was on a panel recently with a Brazilian system leader who was firmly in the camp of standards-only. But when I told him we could measure human skills, his attitude changed completely. Now we had translated these skills into something he could build a system of improvement around. I think this is the most important area that AI can improve schools. Now we can take human skills seriously, allowing curricula focused on human development to be rigorous and intentional. In my opinion, because objective ratings weren’t possible before AI, most Social Emotional Learning (SEL) research was theory. But we can take these hypotheses, try out the methods, measure the results, and figure out what actually works now. And we can study Flourish graduates against matched non-Flourish peers and figure out which human skills really make the biggest difference in high school and beyond. It’s going to be an incredible time of progress for the ways to improve human skills.
Through Dunce Capital, you’ve backed category-defining companies like Outschool and Lambda School. As an investor, you often advise founders to “take the money” even from those who don’t fully understand the space because “money is fuel.” Beyond just capital, what is the one “contrarian” trait you look for in an edtech founder today that signals they are ready to build a system that competes with the status quo rather than just selling into it?
Great question. I started Flourish because I couldn’t answer that question, even for myself. To succeed as a vendor to school systems, you have to build things that they want today. Products like SchoolAI or Magic School are great examples of taking AI and meeting schools where they are today. A product like Project Read’s AI Tutor is still one step ahead of where even early-adopting schools are today. So I wanted to build a school system where we just assumed that AI was going to take over what we used to do, and spend our time and energy figuring out what a modern classroom should focus on, what made a great teacher in the modern classroom, and how to prepare students for the AI age. Without deep use of AI for things like Tier 1 instruction and human skills ratings, that would just not be possible.
I should also say that my main advice to school leaders is to hire their first software engineer. With AI tools, productivity has increased 100x, so you can build things that would have been impossible five years ago. School systems taking back the power to build software that perfectly fits and reinforces their systems is one of the biggest upsides as they work to reshape themselves.
‘…I should also say that my main advice to school leaders is to hire their first software engineer.’
Looking toward the next decade, you’ve predicted that AI will eventually handle the “foundational” work of schooling, freeing humans to focus on mentorship and community. How do you see this evolution reshaping the physical “school day” by 2035, and what is your boldest prediction for how families will navigate their learning choices in that AI-integrated future?
Frankly, my greatest fear is that the average classroom of 2035 looks exactly as it does today. Flourish was created to be a proof point for how this could be done differently, and we hope to accelerate the rate at which larger systems adapt. We are at Day One of our journey, but we are creating what we hope those classrooms will look like for larger systems by 2035. Some time this year, we will open our doors for system leader tours in our Phoenix schools. Rocketship was the poster child for edtech with our learning labs and serious use of technology, and thousands of leaders got their first idea of how to use edtech on our tours. We aim for Flourish to be the place a school leader and their team can go to help them brainstorm what their modern classrooms will look like.
—
Victor Rivero is Executive Producer of Future Focus Forums (F3) and Editor-in-Chief of EdTech Digest. Write to: victor@edtechdigest.com
0 Comments