About the tour

For each of the products we included in Year 1 of our tour, we followed the same process for gathering information.

• We talked with the product team: We started with a founder or product team conversation to understand what the product is, its instructional function, how it's built, and where AI lives in the product design.
• We talked with educators who use it: We interviewed teachers and leaders who had used the product with students for at least three months. We found them through several channels: school system leaders we had relationships with, introductions from product founders and their teams, and our advisory council. We asked the same questions to everyone: what use looked like day to day, what was working, where they felt friction, and what they were seeing in terms of student impact.
• We saw it in action with students: For each product that had a clear instructional use (13/20), we observed it in action with students—what students did, what teachers did, what the AI did, how it landed in the room overall, and how it worked for different groups of students (e.g., students with disabilities, multilingual learners). Where possible during these observations, we talked with teachers and leaders about how they thought about implementation and with students about their impressions from using the product. For most products we observed at least two classrooms. For the seven products that are multifunctional platforms or large language models, we conducted additional user interviews in lieu of classroom visits.

Consistent with our expertise, we focused our observations and analyses on instruction; we did not examine products comprehensively. For example:

• We did not analyze product design for data privacy and student information protection.
• We did not look at how the product integrates technologically with other systems (e.g., student information systems).
• We did not look at pricing because it is changing quickly and hard to track in a standardized way.
• Though we attended to trends in how students, teachers, and leaders thought about it, we did not systematically evaluate the way products promote students’ AI literacy. We focused our review on how the products impacted instruction against current, adopted state standards.
• We did not analyze the environmental impact of specific products.

These—and many other factors—introduce significant concerns and are incredibly important for school systems and everyone invested in K–12 education to consider; however, they are outside the scope of our expertise. We urge organizations with relevant subject matter expertise to analyze products through these lenses and share their conclusions.

We did not choose these products scientifically. We picked these 20 based on products that our partners were using and through a landscape analysis designed to obtain a sample representing a range of grade levels, subjects, and designs.
We focused Year 1 of our tour on student-facing products because we think they have the biggest potential to change the student experience, learning outcomes, and the teacher role. Teacher-facing products (e.g., tools for lesson planning, grading, or getting feedback that are not designed for direct-to-student interaction) can make it easier and faster for teachers to do parts of their complex and challenging job; however, student-facing AI products have the potential to change set up of the teaching job. We wanted to learn at that leading edge.

A sample of 20 products is not a comprehensive review of the full array of available AI-powered instructional products, and our observations may not extrapolate to all schools and school systems (e.g., early adopters may have different strengths and limitations than educators more broadly). It's our hope that a snapshot of the current landscape can support a baseline understanding of AI-powered products and their potential—to help and to hurt—in schooling.

We interviewed teachers and leaders who had used the product with students for at least three months. We found them through several channels: school system and school leaders we had relationships with, introductions from product founders and their teams, and our advisory council. We asked the same questions to everyone so we could see where experiences converged and diverged: what use looked like day to day, what was working, where they felt friction, and what they were seeing in terms of student impact.

For each product, we interviewed teachers and leaders who used the product for at least three months. We asked everyone the same questions. Across those conversations, patterns emerged around what they found useful and where they experienced friction while using the product. We included the values and limitations that the majority of interviewees raised, that were specific to the product rather than true of AI tools generally, and that had implications for instruction and student learning. We did not report on trends in user feedback that were outside our focus on instruction.

We brought the same expectations to AI-powered products that we bring to any instruction—expectations grounded in the evidence base on pedagogy, content pedagogical knowledge, and professional learning. For every product, we followed the same protocol; we:

• debriefed immediately after observing the product in use and recorded values and concerns;

• drafted wins and limitations after completing user interviews, checking each against the observation notes; and

• shared the draft with the product developer, who could flag inaccuracies (developer comments frequently changed the way we worded a limitation, but they never resulted in us removing a limitation—in some cases their pushback sent us back to the evidence and helped us better understand what we saw).

Once 10 profiles were written and reviewed, we ran a cross-cutting analysis using our own coding system alongside rubrics and coding systems generated by Claude and Gemini to surface the criteria emerging across products. Then, we re-reviewed every product against those criteria so that products with similar designs received consistent treatment.

We used AI to support evidence gathering and analyses in circumscribed ways: We built a fieldwork hub to store and synthesize each entry (e.g., notes, meeting transcripts) against a consistent set of categories and used a companion dashboard, built on Lovable and powered by Claude and Gemini, to surface patterns. We used AI as a research partner and to assist with drafting and feedback synthesis, but every word you are reading was carefully chosen by our (human) team.

We've formed an advisory group to provide input and guidance across every stage of this project—from who we are learning from to how we share our findings. We chose the advisors to reflect a range of perspectives and to help us understand our blindspots—priorititizing practitioners (e.g., teachers, school leaders, and system leaders). Advisors helped us understand the range of products at the beginning of our process, reviewed the trends and takeaways we identified to give us feedback on substance and clarity, and reviewed website design. They provided feedback on the structure of the product profiles, but they did not review or influence the content for any of the product profiles.

• Auditi Chakravarty, Advanced Education Research and Development Fund
• Aaron Cuny, AI for Equity
• Courtney Allison, Ed Reports
• Emma Doggett, aiEDU
• Trenace Dorsey-Hollins, Parent Shield
• Jennie Dougherty, KIPP Northern California
• Buffy Edelstein, Frisco ISD
• Heather Gauck, Grand Rapids Public Schools
• Kristen Hole, Houston ISD
• Laurence Holt, Advisor
• Luke Kohlmoos, Accelerate (Instruction Partners cofounder)
• Robin Lake, CRPE
• Jeff Livingston, EdSolutions
• Sarah Lydick, Hamilton County Schools
• Emmalie Moseley, Denver Public Schools
• Karen Nussle, Ripple Communications
• Kira Orange-Jones, Teach Plus (Instruction Partners board member)
• Alejandro Ramos, Clint ISD
• Kali Peracchia, Denver Public Schools
• Tanji Reed-Marshall, Liaison Educational Partners (Instruction Partners board member)
• Michelle Rhee, EO Ventures
• Sam Ribnick, A-Street Advisors
• Kevin Shaw, KIPP New Jersey
• D'Andre Weaver, Digital Promise
• Joanne Weiss, Consultant and Former Chief of Staff at U.S. Department of Education (Instruction Partners board member)
• Mary Wells, Bellwether (Instruction Partners board member)
• Caitrin Wright, Silicon Schools Fund

In addition to our advisory group, we are deeply grateful to the following people who shared their insights and perspectives with a learning spirit and made this tour and our ability to share the findings possible:

• The students, teachers and leaders that welcomed us into your classrooms so we could see products in action. Thank you to the educators of Loudon County VA, Houston ISD, North Bellmore, Manhattan Business Academy, Denver Public Schools, Jackson MS, Dream Charter Schools, KIPP Academy-Northern California, KIPP-Hatch Academy, KIPP-Purpose Academy and KIPP-Upper Roosevelt Academy, Kindle Education, and NYCPS District 11.
• The teachers and leaders who spent time with us on calls sharing your experience with products. We promised you anonymity, and we will honor that, but you know who you are and we are grateful for your insights.
• The product developers who taught us so much, thank you for sharing your philosophy, product design approach, insights on data, and, in some cases, your networks, and thank you for working with us to tell the story in a way that is honest, transparent, and fair.
• The members of the Instruction Partners team put in miles of travel, hours of interviews, and so many rounds of editing and sweat over every design decision. This content was chiefly written by Emily Freitag and Doe Kim with support from across the team, including substantive review by the entire leadership team.

As a nonprofit organization, we do not hold equity or any financial interests in any of the companies behind the products featured on this tour. We were connected to product developers through school systems we work with and through direct outreach.

Currently, Instruction Partners works with a number of schools that use Coursemojo, and we work with the company directly to support product implementation alongside school leaders. This work is funded by grants and district contracts, no funding comes from the product company. We will be working with several schools using many of these products in the year to come and we will disclose those relationships on each product page if/when those partnerships are confirmed.

Our AI Learning Tour team is NOT the same team that supports product implementation in schools (i.e., no staff who contributed to this project supported any of our work with products in schools nor will they be involved in any future product implementation support work). No user interviews included in this tour were conducted with employees of schools we work with.

The costs of Year 1 of the tour were funded by an unrestricted grant from MacKenzie Scott awarded in 2022 and dedicated grants from Silicon Schools Fund and Sue Lehmann. We are deeply grateful for the support that made the learning and sharing possible. We do not engage in any funding relationships for the tour that create conditions on which products we include (or exclude) or allow funders to comment on the product-specific information we release.

We support instructional leaders. If AI-powered instructional tools have a beneficial role to play in supporting student learning, we suspect that leaders’ decisions about these products (e.g., selection and implementation) will factor into how effectively they lead teaching and learning. We want to help our partners—schools, school systems, and states—navigate the emerging options.

The bulk of our work (i.e., 80%+ of our work) continues to focus on helping leaders improve instruction through implementation of high-quality instructional materials in paper-pencil form.

This tour is a three-year project (the publication of this website marks the end of Year 1).

After three years, we think that the sector will transition from “early stage sense making about the potential role of AI-powered instructional tools” to “clear market signals about quality needed.” At that point, we think an organization built for the purpose of reviewing products and materials will be required to continue this work—and Instruction Partners has no intention of serving that function. Our focus will continue to be on helping school, school system, and state leaders select and implement high-quality instructional materials.

We're excited to continue our learning on this tour—and continue sharing what we learn along the way. In the 2026–27 school year (Year 2 of the tour), we plan to:

• Study and see in action more student-facing AI-powered instructional tools that are used in more than 10 school systems (we expect to add another 20–40 products).
• Expand the tour to include 10–20 teacher-facing products.
• Partner with Stanford University to conduct a secondary analysis of the existing independent studies product developers cite as evidence of impact.

If you have feedback or questions about the insights shared on this website, or if there is a student-facing AI-powered product that you think we should study in the next phase of our tour, please email communications@instructionpartners.org and a member of our team will be in touch.