SchoolAI works with students. That shapes every product decision we make: what we build, what we won't build, and how we hold ourselves accountable.
Explore our principles, compliance certifications, privacy documentation, and impact evidence. We publish them because effective AI in education is built on public trust.
The Every Student Succeeds Act (ESSA) defines four tiers of evidence for educational interventions. SchoolAI has earned two ESSA Level III “Promising Evidence” designations on independently validated two-year studies, reviewed by Instructure's EdTech Collective — a third-party research organization with WWC-certified reviewers.
A cross-sectional study of 390 educators across 48 U.S. states examined whether greater experience using SchoolAI predicted better teacher outcomes. It did, significantly, across all three measures: perceived productivity and time saved, instructional effectiveness, and professional well-being. Effect sizes (d = 0.30–0.42), or roughly equivalent to moving a teacher from the 50th to the 62nd percentile on these measures, with controls for years of teaching experience.
A two-year implementation study examined whether teacher engagement with SchoolAI predicted higher levels of critical thinking engagement, as measured through conversations. Mean critical thinking scores rose 28.3% from October 2023 (M = 1.38, SD = 0.49) to October 2025 (M = 1.77, SD = 0.69). Levels 3 and 4, the deepest levels of analytical and evaluative thinking, more than doubled. The study tracked 82 teachers and 13,882 student-AI conversations in a U.S. suburban public school district.
Dot's instructional behavior is based on what the research says produces durable learning. It is built on how brains learn and how good teaching supports them.
Our Theory of Change maps how our mission, pedagogy, and product design connect to measurable student, teacher, and administrator outcomes.
Cognitive Load Theory
Working memory has hard limits, and overload prevents learning. Dot calibrates response complexity by grade level: vocabulary, sentence length, and the number of concepts introduced at once.
Sweller, 1988; Sweller, van Merriënboer & Paas, 1998
Warm Demander Pedagogy
High warmth and high expectations held simultaneously. Dot is patient and encouraging without lowering the bar. Validate the student's experience, then return them to the work.
Kleinfeld, 1975; Delpit, 2012; Hammond, 2015
Schema Theory & Prior Knowledge
New learning is integrated into mental frameworks built from prior knowledge — and not added to an empty slate. Dot activates what students already know before introducing new concepts and connects new material to existing understanding.
Bartlett, 1932; Ausubel, 1968; Anderson, 1977
Zone of Proximal Development & Scaffolding
Learning happens at the edge of what a student can do alone with support. Dot scaffolds through progressively narrower questions and only models a step when scaffolding has run its course.
Vygotsky, 1978; Wood, Bruner & Ross, 1976
Retrieval Practice
Pulling information from memory strengthens it more than re-reading or re-watching. Dot's default is to ask students to recall and explain rather than receive information passively.
Roediger & Karpicke, 2006; Dunlosky et al., 2013
Socratic Dialogue
Guided questioning develops reasoning more reliably than direct explanation alone. Dot asks "What's your first thought?" before offering a menu of approaches.
Paul & Elder, 2007; Chi et al., 2001
Spaced Repetition & Interleaving
Distributing practice over time and mixing problem types produces more durable learning than cramming. Dot's design supports spacing and interleaving in how it sequences questions and revisits prior concepts.
Cepeda et al., 2006; Rohrer & Taylor, 2007
Formative Feedback
Specific, in-the-moment feedback produces larger learning gains than evaluative praise after the fact. Dot names where reasoning broke down rather than handing over a correct answer.
Black & Wiliam, 1998; Hattie & Timperley, 2007
Desirable Difficulties
Learning that feels harder in the moment often produces stronger long-term retention. Dot introduces productive struggle just beyond the student's current level rather than smoothing friction away.
Bjork & Bjork, 2011
Metacognition
Students learn more deeply when they monitor and articulate their own thinking. Dot asks students to walk through their process and name where things stopped making sense.
Flavell, 1979; NRC, How People Learn, 2000
Science of Reading
Reading is a learned skill that depends on systematic, explicit instruction in phonemic awareness, phonics, fluency, vocabulary, and comprehension. Dot aligns with this evidence base and redirects requests built on unsupported approaches like three-cueing.
National Reading Panel, 2000; Seidenberg, 2017
Standards Alignment & Outcome-Driven Design
Effective instructional design starts from defined learning outcomes and works backward. SchoolAI maps Dot to state and national standards and evaluates impact against measurable outcomes.
Wiggins & McTighe, 2005
Cognitive De-skilling / Up-skilling
AI can remove the cognitive work that produces learning, and that work is where neural pathways form. SchoolAI is designed to preserve that friction and channel it, building student capability through AI rather than outsourcing capability to AI.
Van Damme & Fadel, 2026
Universal Design for Learning
Good teaching designs for learner variability and cultivates agency: purposeful, reflective, resourceful, authentic, strategic, action-oriented learners. It is access rooted in the student's capacity to direct their own learning. Dot adapts vocabulary, scaffolding depth, and visual/auditory entry points based on what each student is showing.
CAST UDL Guidelines 3.0, 2024
We believe EdTech companies have a responsibility to contribute to the policy and research ecosystem, not just draw from it. Members of our team participate in external governance, advisory, and legislative efforts related to AI in education.
We're actively seeking research partnerships with universities, districts, and evaluation organizations to advance our evidence-building pipeline. If you're a researcher interested in studying AI's impact on student critical thinking, teacher effectiveness, or the cognitive effects of AI-mediated learning, we'd like to hear from you.
