States and school systems considering the future of assessment find themselves in a position somewhat analogous to Netflix circa 2007. Netflix had a thriving DVD-by-mail service. It kept that approach while simultaneously embracing the future by introducing streaming—adding personalization, usefulness, and usability without abandoning its core business.
For Pre-K12 education systems, the path forward requires a similar “yes, and” approach: building upon what we have, while embracing the future. We must maintain high expectations, instrument quality, and transparency (the “DVDs”) while investing in the “streaming” equivalent of assessment—innovations that are continuous, adaptive, and instructionally useful.
Innovating the What, How, and For Whom
To achieve this, we must innovate the “What,” “How,” and “For Whom” of assessment systems.
- What: Measure what matters most, not just what’s easy to measure. This means moving from single-score verdicts to balanced assessment systems; building upon ambitious grade-level teaching while pivoting toward instructionally useful assessment and practical measurement for improvement.
- How: Leverage AI for rich insights. Transition from one-size-fits-all models to customized approaches powered by Safe, Accountable, Fair, and Effective use of technology to achieve person-specific insights that honor learner variation.
- For Whom: Cultivate “assessment-capable learners“—students who can interpret and act on feedback to take charge of their own improvement—and educators who need evidence to inform the why and how of daily instruction.
An Instructive Engine of Innovation: ED/IES SBIR
This “streaming” future is not hypothetical; the building blocks are already being developed through strategic public investment. For two decades, for example, the U.S. Department of Education’s Small Business Innovation Research program (ED/IES SBIR) has served as the federal engine for seeding research-based, commercially ready education technology.
The program funds R&D and rigorous evaluation through competitive, milestone-based contracts. Crucially, developers must partner with researchers and schools to ensure innovations are iteratively refined, scalable, and sustainable.
The return on such public investment is profound. Our report, From Seed Funding to Scale, shows significant returns: $91 million in R&D funding reached over 130 million users (approx. $0.70/user) and generated $828 million in returns, a 9:1 return on investment.
Over the past 20 years, ED/IES SBIR has supported products that reimagine how students learn and educators teach. The examples highlighted in this article are proof points that investment in innovation matters. They represent potential features of future assessment systems, demonstrating how states might begin to safely and effectively embrace assessment innovation without pre-emptively dismantling the present.
The New Era of Assessment: Emerging Themes from the Portfolio
The ED/IES SBIR portfolio offers more than just isolated success stories; it provides some beginnings of a taxonomy for innovation. The products emerging from these investments illustrate precisely how novel solutions (often adjacent to but suggestive of AI-driven assessment) usher in new approaches to learning. These innovations facilitate experiences that are personalized, continuous, active, and embedded in authentic learning.
Across subjects and age levels these technologies demonstrate emerging themes worth considering.
1. Personalized and Adaptive Learning in Real Time
AI enables assessment and instruction to occur simultaneously. By analyzing student artifacts and learning process data, these solutions provide more formative feedback that guides individual learning progressions.
StepWise, for example, acts as a virtual math tutor. Instead of grading a final answer, it uses AI to analyze the student’s specific steps and logic, identifying moments of confusion and delivering individualized hints to get them back on track. Similarly, Capti personalizes literacy assessments, recommending specific supports based on diagnostic results. Meanwhile, tools like PACE AI are converting context-specific materials into personalized English-learning lessons that adapt as the learner progresses.
2. Beyond the Screen: Active, Hands-On, and Collaborative
Contrary to the myth that AI means more screen time, multi-modal AI extends learning into tactile, social, and exploratory experiences.
We see this in KASI, which uses computer vision to help students—including those with visual impairments—learn chemistry by manipulating physical, tactile molecular models. The AI “watches” their hands and provides audio feedback. PocketLab embeds AI feedback directly into wireless sensors, allowing students to conduct physical science experiments while the system works to sharpen their inquiry skills.
For early learners, Kibeam and SoundTown turn reading and phonics into interactive play using screenless devices. A child reads a physical book or speaks to a character, and the device uses Natural Language Processing to listen and support them, keeping their eyes on the text or the task, not a tablet.
This active approach extends to authentic professional simulations. INQits and 2 Sigma Schools immerse learners in virtual science labs and coding environments that mirror real-world complexity, assessing students on how they design investigations or debug code, not just on what they recall.
3. Learning Integrated Assessment
Assessment need no longer be a separate or final event to categorize students after the fact; it can be integrated into learning itself.
Writing assessment has traditionally been labor-intensive and delayed. Tools like Scrible, CG Scholar, and Revision Assistant (integrated into Turnitin) embed assessment directly into the composition process. Students receive “Signal Checks” on their evidence, structure, and argumentation while they are writing, allowing feedback to serve as a scaffold for improvement rather than just a final judgment. Similarly, Moby.Read uses AI to analyze oral reading fluency, instantly diagnosing accuracy, rate, and prosody (expression) to help educators tailor literacy interventions without the delay of manual scoring.
SownToGrow measures social and emotional well-being. Using machine learning to analyze student reflections, it helps educators identify trends in student emotional health, flagging those at risk so schools can intervene with support rather than discipline.
4. Dashboards with Insights for Educators
Dashboards transform data into real-time insights, shifting educators from didactic delivery to targeted support.
OKO, a collaborative math game, provides a dashboard that lets teachers see how groups are communicating and problem-solving, enabling immediate intervention if a team gets stuck or conflicts arise. Meanwhile, Education Modified tracks IEP progress and centralizes goals, providing special education teams with the guidance they need to ensure no student falls through the cracks.
The Pathway Forward
These four themes do not constitute a comprehensive set of features for quality assessment innovation. However, these examples suggest that the technology required to upgrade state, school system, and classroom assessment systems is a present reality, validated by rigorous research and user testing. The “DVD” era of standardized testing provided us with a baseline of data. The “Streaming” era, powered by such innovations, offers us a pathway to usefulness, personalization, and student learning.
This blog series on Advancing AI, Measurement and Assessment System Innovation is curated by The Study Group, a non-profit organization. The Study Group exists to advance the best of artificial intelligence, assessment, and data practice, technology, and policy and uncover future design needs and opportunities for educational and workforce systems.
The post Education’s Netflix Moment: A “Yes, And” Approach for Assessment System Innovation appeared first on Getting Smart.