Last month Accelerate announced the eleven grantees selected for our 2026-27 Call for Effective Technology (CET). Year one of our CET program is still wrapping up — final research reports are expected later this summer — but its lessons are already shaping our approach to grantmaking. This incoming cohort reflects some of what we are learning, and offers an early read on where K-12 AI-powered tools are headed.
We received 87 applications for our 2026-27 CET program and selected 11 grantees: seven focused on math, three on literacy, and one covering both. The tools serve students from PK through grade 12, and our grantees together operate across 45 states.
- DREAM — an AI tutoring assistant that synthesizes student data and curriculum to generate targeted lesson plans and pre-teaching content for human tutors working with students two or more grade levels behind
- EdLight — an AI formative assessment platform for grades 3–12 math that captures and analyzes student work, including handwriting on paper, and surfaces real-time insights for teachers
- Eedi — a K–12 math platform built around 60,000+ diagnostic questions that map student errors to specific misconceptions, paired with AI tutoring grounded in those misconceptions
- eKidz — a grades 1–6 literacy platform that uses AI voice recognition to assess oral reading fluency and a constrained AI tutor to scaffold writing revision, with multilingual support
- EVIDENCEB — an adaptive K–5 math platform (Adaptiv’Math) that uses reinforcement learning to personalize learning pathways in real time, grounded in cognitive science research
- Grokkoli — a PreK–5 math tutor built first for students with IEPs and others with significant gaps; when a student gets stuck, it breaks problems down, switches representations, and fades support as mastery emerges
- Inletech (ArtMath) — a grades 3–6 math platform that uses Vision-Language Models to scaffold visual problem-solving and lets students co-create math stories that deepen conceptual understanding
- Learnie (Thinkverse) — an AI math co-teacher for grades 3–12 that guides students through standards-aligned practice with graduated hints, misconception-aware prompts, and prerequisite refreshers across 100+ languages
- Lit (Bloom) — a K–2 foundational literacy platform that uses speech recognition for daily reading practice and surfaces real-time coaching insights for teachers
- Littera Education — a grades 3–8 math tutoring model that pairs Littera’s small-group human virtual tutoring with Third Space Learning’s Skye AI tutor to deliver high-dosage tutoring at lower cost
- ROYO.AI — a K–3 literacy platform that generates personalized decodable books at each student’s exact phonemic level, with AI voice recognition providing in-the-moment feedback as students read aloud
Several through-lines across the 2026-27 cohort reflect what our first cycle of CET is teaching us.
Stronger evidence designs. Our first set of CET evaluations was almost entirely quasi-experimental. That gave us a starting evidence base, but it has limits. For cohort two, we expect several grantees to run full randomized controlled trials, paired with lighter, faster studies in others. The result is a more layered evidence base: studies that can speak with confidence about effects, alongside more practical learnings about implementation, fit, and feasibility.
More students per implementation. We raised the minimum student count for this cycle. Larger samples improve generalizability and let us ask more honest questions about sub-groups (i.e., we can say with increasing certainty which students benefit most from which tools).
Personalization aimed at very specific gaps. The tools in our second CET cohort are not trying to be everything to everyone. Several are built specifically for students with IEPs or others with significant gaps. Others target practice at exactly the phonemic or skill level a student needs. Many align with Tier 1 instruction — reinforcing, rather than replacing, the core curriculum. The tools, increasingly, can target specific student gaps with a resolution that wasn’t feasible a few years ago.
Accessible by design. Year one of CET taught us that students engage most with tools that let them speak, draw, and create. Tools in this cohort continue to deliver novel uses of the technology. One reads handwriting on paper and turns it into actionable insights for teachers. Many include text-to-speech and speech-to-text. Several operate in multiple languages. Accessibility, in this cohort, is not a feature but the price of entry.
AI and humans, together. A few 2026-27 grantees signal where the future of AI and tutoring might be headed. One is a deliberate pairing of human tutors and AI support, designed to bring high-dosage tutoring to more students at lower cost. Another built an AI assistant for human tutors and school-based personnel; the school’s own high-quality instructional materials live inside the tool, which also pulls in student assessment data down to a student’s specific skill gaps. It generates coherent, targeted, pre-teaching content for each student, in near real time.Our first CET grant cycle taught us two things: implementation of AI tools varies widely, shaped by teacher decisions about what to use and why — and rigorous evaluation of these tools is possible. The 2026-27 CET cohort is an opportunity to strengthen our research designs, reach more students, and evaluate tools that have sharp focus on solving specific problems in K-12. We are grateful to these eleven teams for opening their tools to research and helping build the evidence base students deserve.
Our first CET grant cycle taught us two things: implementation of AI tools varies widely, shaped by teacher decisions about what to use and why — and rigorous evaluation of these tools is possible. The 2026-27 CET cohort is an opportunity to strengthen our research designs, reach more students, and evaluate tools that have sharp focus on solving specific problems in K-12. We are grateful to these eleven teams for opening their tools to research and helping build the evidence base students deserve.
Jennifer Bronson is the Managing Director of Learning Programs at Accelerate.
