Colleague AI wins NSF Phase II grant to expand K-12 AI tools

Colleague AI, a University of Washington spinoff focused on K-12 education, received a National Science Foundation SBIR Phase II grant worth $1.25 million in base funding and about $2 million with expected supplements. The money will fund new agentic AI features, deeper school-district research, and expanded testing across partner districts over the next 18 to 24 months.

Why it matters: - The grant gives Colleague AI more runway to build K-12-specific AI tools designed for teachers, students and district leaders. - The funding also signals continued federal support for education AI tied to research, co-design and school-based testing. - Colleague AI expects the work to strengthen product quality and adoption in schools that want specialized AI features.

What happened: - Colleague AI won a Phase II Small Business Innovation Research grant from the National Science Foundation. - The award includes $1.25 million in base funding. - Supplemental funding is expected to bring the total to about $2 million. - The grant period runs for the next 18 to 24 months. - The company announced the award on Aug. 19, 2026.

The details: - The Phase II award follows Colleague AI’s Phase I SBIR/STTR project. - Phase I funded four specialized AI models for lesson quality measurement, retrieval and augmentation, instructional nudging and subject-specific content generation. - Those models now power the company’s lesson-planning and grading tools for K-12 teachers. - New funding will support research, development, testing and release of major platform additions. - Planned additions include finely tuned AI assessment tools, student feedback systems and an on-demand classroom-data AI analyst. - Colleague AI also plans three-actor learning tools for teacher-student-AI experiences. - The company is developing advanced student AI tutoring with deep memory and context about subjects, curriculum, student competencies and specific classroom assignments. - The grant will also support new features in grading, tutoring, student discussions, administrative tools and actionable data insights. - Colleague AI will use collaborative research, co-design with K-12 stakeholders, extensive evaluation and testing with school-district alpha and beta partners. - The company is seeking district partners with feature requests and offers up to $250,000 in additional funds for co-development. - Colleague AI has signed MOUs with 10 school districts to study instructional practice and student learning outcomes during the 2026-27 school year. - The company says it will extend new platform additions to partner districts and Pro subscribers as they become ready. - Product announcements are expected later this year. - Colleague AI says it now serves more than 250,000 users nationwide. - The platform is available to schools, districts and other K-12 learning organizations.

Between the lines: - The emphasis on agentic AI and custom data architecture suggests Colleague AI is betting on deeper workflow automation, not just generic classroom chatbots. - The district studies and co-design model point to a strategy that blends product development with evidence-building, which can matter to school buyers focused on safety and instructional quality. - The company’s pitch centers on research-backed tools that outperform competitors on quality, but the strongest near-term proof will come from school results and district uptake.

What's next: - Colleague AI has already started building new features under the grant. - The company will roll out platform additions to partner districts and Pro subscribers as they are ready. - District research partnerships are expected to continue through the 2026-27 school year. - Additional product announcements are due later in 2026. - Schools and districts interested in participation can contact Colleague AI directly.

The bottom line: - The NSF award gives Colleague AI fresh funding and validation as it pushes from core lesson-planning tools into broader AI infrastructure for K-12 schools.