Today, Kiddom announced Atlas, a new AI-enabled Learning Intelligence Technology embedded in Illustrative Mathematics®. We’re proud to share the role ANet played in building the curriculum-aligned assessment engine that powers it.
For years, educators have been asked to work across disconnected systems — curriculum in one place, assessments in another, intervention somewhere else entirely. The burden of making those systems talk to each other has fallen on teachers.
Atlas starts with a different question: What would it look like if assessment were designed to live inside instruction and actually power it?
That question is what brought Achievement Network (ANet), Kiddom, and Teaching Lab together, and it’s what sits at the heart of Atlas.
As AI becomes more present in classrooms, the volume of instructional activity, practice, and data is increasing rapidly. But more activity doesn’t automatically mean more learning.
Without trustworthy learning signals, AI risks amplifying the illusion of progress rather than progress itself.
Our work centers on High-Quality Instructional Assessment (HQIA): the idea that assessment should be part of instruction itself, not a separate system that reports results after the fact. HQIA is intentionally designed to surface student misconceptions in real time, tied directly to what students are learning, so teachers can respond immediately inside daily instruction.
In Atlas, that meant rethinking assessment from the ground up, not as a checkpoint but as infrastructure.
Over our 20-year history, ANet has built one of the nation’s most granular datasets on student math misconceptions aligned to high-quality instructional materials. That expertise informed our development of daily cooldown items embedded in Atlas, intentionally designed to surface misconceptions and generate reliable learning signals.
Each cooldown was crafted to:
We applied ANet’s longstanding commitment to item quality, alignment to curriculum and standards, and carefully constructed distractors to these daily lesson cooldowns. That meant adapting production processes and expanding capacity while maintaining the alignment, rigor, and reliability that make high-quality assessment signals trustworthy.
When assessments are aligned directly with instruction, insight and action travel together.
In Atlas, patterns from formative cool-down questions help make student thinking visible while teachers are still teaching the content. Because Atlas lives inside the curriculum platform educators already use, those signals do not require a separate dashboard, login, or workflow.
Rather than pulling students out of grade-level work or sending teachers to disconnected tools, assessment signals support:
This reflects a shared belief across partners: differentiation should happen inside grade-level instruction, not apart from it.
Atlas is not about adding another tool to an already crowded ecosystem. It’s about addressing a long-standing implementation problem districts already feel: fragmentation.
By intentionally designing curriculum, assessment, and instructional decision-making to function as one system, Atlas offers a concrete example of what instructional coherence can look like in practice – supported by AI, but grounded in learning science and disciplined assessment design.
For ANet, this work also represents a broader opportunity: reaching students and teachers through partnerships that bring high-quality assessments directly into platforms educators already use, while preserving the rigor and purpose that make assessment meaningful.
Atlas is still early, and we are learning alongside our partners. But it offers a glimpse of what’s possible when assessment is treated not as a checkpoint, but as infrastructure that allows teaching and learning to move forward together.
We’re excited to continue this work with Kiddom and Teaching Lab, and to keep learning from teachers as this approach reaches more classrooms.