Colleague AI

06/01/2026

One of the things educators LOVE about Colleague AI's platform is that all of the interactive activities and assignments generated are accessible - friendly to learners who need accommodations.

Just click the accessibility man and a host of options appear!

No other platform can generate interactive assignments and activities in 5 minutes, let alone make them accessible.

05/29/2026

The Center for Reinventing Public Education wrote an excellent article recently discussing what most K-12 AI tools or platforms miss:
https://crpe.org/getting-beyond-the-lightbulb-stage-why-ai-is-not-yet-transforming-education/

Some highlights:

Most AI tools are standalone and serve 1 function. Colleague AI is not.

Most AI tools have a "weak grounding in learning science" - this is why we spent years with education researchers and learning science professionals building Colleague AI - BECAUSE PEDAGOGY MATTERS!

Most tools are not aligned with education’s real challenges. Teacher productivity cannot be the goal in order for AI to have a positive impact. Colleague AI asks "what are your districts strategic priorities" and then we help you customize the platform accordingly and adjust training accordingly.

Not all AI tools are equal.

New AI tools are entering the ed tech market every day. But how are they actually playing out in schools and classrooms? Drawing on semi-structured interviews with more than 50 stakeholders across ed tech, philanthropy, policy, teaching, and advocacy, this brief identifies the gaps between what AI d...

05/26/2026

Here's a real thing that happens when you train teachers on Claude: you hit the credit limit mid-session and get locked out for five hours. In the middle of a 4–6 hour professional development day. With a room full of educators who just got excited about AI.

Colleague AI was purpose-built for K-12 — not adapted, not retrofitted.

It already has agents designed for educators, no timed usage limits that can derail a training session, and it's priced for district budgets, not enterprise contracts.

When you're investing in teacher PD, the last thing you need is the platform tapping out before lunch.

05/22/2026

Schools don't run on one job title. They run on the teacher who stays late, the counselor who notices, the custodian who has the building ready before sunrise, the bus driver who knows every kid's name, the front office that fields a hundred questions before 9am, and the leaders holding it all together.

To all of you — and everyone in between — thank you. Rest this long weekend. Recharge. You've earned it.

05/21/2026

Our latest research is now live on arXiv: "Generative AI in K-12 Classrooms: A Midyear Implementation Report."

Produced jointly by Colleague AI and AmplifyLearn.AI at the University of Washington, the report looks at how teachers actually used AI across 12 school districts during the first half of the 2025–26 school year. These weren't cherry-picked pilots — the districts range from a few thousand students to thirty thousand, and span rural, suburban, and urban communities. The analysis combines platform usage data with district-provided administrative records to move past anecdote and into evidence.

This is the kind of work the field needs more of: transparent, mid-implementation data on what teacher AI adoption really looks like at scale. We're sharing it openly because districts deserve to make decisions grounded in evidence, not vendor claims.

Read the full report: https://arxiv.org/abs/2605.16277

05/20/2026

New video podcast episode out now: Research-Empowered Education Technology: Applying Theory to Practice.

Host Ryan Fries, an educator in Bellingham, WA, sits down with Michael Xiao, an education researcher at the University of Washington's national AI in education research center AmplifyLearn, to dig into one of the hardest questions in K-12 right now: how do you take what the research actually says about learning — and put it into the tools teachers use Monday morning?

Check out our snippet, and watch the entire podcast here: https://www.youtube.com/watch?v=zjDc6KR-X4o

05/19/2026

Last week, our CEO Dr. Min Sun joined a panel hosted by the League of Education Voters alongside:
- high school students
- Jolenta Coleman-Bush, Director at Microsoft Philanthropies
- Bree Dusseault, Managing Director at the Center for Reinventing Public Education
- Babak Mostaghimi, Founding Partner at LearnerStudio
and others!

The conversation tackled three questions every district is wrestling with right now:
- how AI is actually being used in classrooms today
- what guardrails need to be in place to protect students and families
- what skills students will need to be ready for an AI-shaped workforce.

The student voices in the room made it sharper — they were honest about what's working, what isn't, and what they want from the adults building these tools.

Watch the recording: https://www.youtube.com/watch?v=ms5wN_gr9so

Artificial Intelligence (AI) is here to stay. As this technology continues to evolve, educators must prepare students to understand and use these tools. In W...

05/18/2026

Did you know? Every lesson plan our platform generates is evaluated against quality criteria drawn from the peer-reviewed research on lesson plan quality — clarity of objectives, alignment, coherence, scaffolding, assessment fit, and more.

If a draft doesn't meet the bar, it's sent back to our agents to revise until it does. You don't see the misses. You see plans worth your time.

Then you can go further. Want to strengthen inquiry-based learning, skill development, or another quality dimension for your specific class goals? Run an additional check. The platform surfaces the original text, the proposed improvement, and an explanation of why the change strengthens the lesson — so the decision stays yours.

The teacher is the editor, not the Quality Assurance layer!

05/15/2026

Check out our 60+ questions every K-12 district should ask AI vendors (including us). This is the most extensive list of questions for AI vendors in K-12 today.

Sections cover pedagogy, AI model and customization, FERPA/COPPA/state privacy, equity audits, accessibility (VPAT), district leadership workflows, and implementation depth. Drawn from peer-reviewed research and real district AI addenda.

If a vendor can't answer most of these clearly, that's a yellow, orange, or red flag.

Read the full list:

Put together from academic papers, research, evidence-based practices, and implementation experience, check out the 60+ questions to ask K-12 AI vendors.

05/14/2026

Dr. Lief Esbenshade, one of our researchers with AmplifyLearn, will be presenting at the Stanford Education Data Science Conference on Tuesday, May 27.

His talk, "Midyear Evidence on Teacher's Use of Generative AI in K-12 Schools," brings something the field badly needs: actual data on how K-12 teachers are using generative AI midway through the school year — what's sticking, what isn't, and how patterns vary across context, subject, and experience.

Districts deserve to make AI decisions based on evidence, not vendor claims. Sharing this work at Stanford is part of how we hold ourselves to that standard.

Check out the conference: https://edsconference.stanford.edu/

The Education Data Science Conference 2026 is a two-day forum for researchers, practitioners, and students to shape the future of education through data. We invite submissions that advance rigorous, ethical, and interdisciplinary approaches through using data in a variety of learning contexts. From....