Adaptive Learning: Lessons from Two Decades in Diverse Classrooms

Introduction: The Evolution of my Personalized Approach

I taught for close to 20 years, in four different states. I spent my first seven years in charter and private schools and the remainder of my career in district schools. Not only did I have quite a bit of cultural adaptation to make between schooling models, but I also had a number of regional cultural changes to make. For example, in my second year teaching I learned about the three most common learning styles: visual, auditory, and hands on. For years, I began each year having students do a learning style inventory and gave them learning tips that capitalized on their strengths and supported their challenges. As well, I started teaching in Tucson, AZ. I would do a lot of work with Southwestern desert culture centered metaphors and literature. However, we next moved to Portland, OR, as my wife built her career. Portland’s persistent gray and rain created a different classroom dynamic. The constant drizzle meant more indoor time, more introspective work. Students who had never seen a desert couldn’t relate to my Southwestern metaphors, but they understood Ursula K. Le Guin’s stories of forests and mist. The short winter days and long stretches without sunlight taught me new ways to maintain energy and engagement. A large part of any teacher’s job is to help bring energy to the classroom and to keep everyone focused and motivated. This became an ongoing and daily challenge as I was struggling with my own energy and enthusiasm in the dark and cold.

However, this wasn’t the only lesson in adaptive learning that I took from teaching in Portland. The school that I taught in was a special program wherein the teachers created curriculum that was tailored specifically to the individual student. While this isn’t the type of technological adaptive learning that AIs and other Educational Technology apps use, it’s still the same principle, it’s more of a traditional type of adaptive learning that teachers have always done. Modern, technological, adaptive learning is an educational approach that tailors learning experiences to individual needs and paces, offering personalized instruction and feedback. It uses data and algorithms to assess a learner’s strengths and weaknesses, adjusting the difficulty, content, and pace of instruction accordingly. This method contrasts with traditional “one-size-fits-all” learning, aiming for more effective and engaging learning experiences. At one point, while teaching in Santa Fe, NM my supervising administrator, in an evaluation, stated that what she saw was the power of my teaching is that I didn’t just modify for students with learning plans, like special education or English Language Learners; but that I modify for all of my students.

I think that I’d always been primed for adaptive teaching due to my radically dynamic upbringing. I was often faced with large upheavals that demanded that I adapt to very different living situations. So, taking an adaptive approach to teaching really felt intuitively natural to me. Even before teaching in Tucson, due to the nature of the charter schools that I taught in I mostly had to bring myself up in the profession. Through the use of my intellect, creativity, compassion, and calling to teach, I was always looking for better ways to reach my students, both emotionally and academically.

Once I was teaching in a district school, in Santa Fe, and found an administrator who recognized what I was bringing to the table and helped me to understand that it was not only uncommon but highly valuable I began operationalizing my adaptations. For example, the only position that I’d been able to get over the phone, from OR to NM, was teaching seventh grade math. I had a math endorsement because in OR middle school teachers need to be endorsed in all four core subjects. However, math is, by far, my least strong subject. Fortunately for me the school that I taught at had been taken over by the state and I had a canned curriculum that walked me through, day by day, everything that I had to teach. After this year, when I was able to teach English again it occurred to me that I’d learned a solution I’d been seeking for years — how to better teach writing. Having done my schooling in the 1980s I’m a product of the “Whole Reading” model. We did not learn how sentences are constructed or grammar rules. We were exposed to exemplary writing with the philosophy that we’d intuitively understand the rules and mechanics. Unfortunately, that did not help me when it came to teaching writing. But, teaching pre-Algebra, with the order of operations, put me on the look out for a similar order of operations for writing, which I found.

Because I’d always taught in traditionally underserved schools most, if not all, of my students had had a hodgepodge of strategies taught to them for how to write effectively. My model gave them a strong and concrete foundation from which to use their other writing strategies and from which to build on. What I quickly learned though, is that most students struggled, sometimes for the whole year, with my basic essay, whereas others, at various points, mastered it. This is one of the eternal challenges of educators. My response was to put each student who mastered the basic essay on an individualized learning plan for essays. I’d create (in the pre-learning management system days) lessons that taught them the more sophisticated essays and would offer individualized tutoring if needed. Once they mastered this essay I’d give them a more sophisticated one and would take them as far as they wanted to go.

Practical Applications for Instructional Design

So, now that we are in the days of the learning management systems (LMS) that offer more individualized learning models as well as algorithm supported learning that can adapt to a student’s needs the question is: is there a place for traditional teaching tools that are informed by human insight gained through experience? If you’ve already read my essays about using AI as a collaborative tool, anytime I talk about the synergy of quantitative and qualitative data, or about metacognition you know what my answer is.

The answer is, of course. My manual progression system for essays — moving students from basic to increasingly sophisticated structures as they demonstrated mastery — operated on principles similar to what platforms like ALEKS now automate. Where I once maintained mental maps of each student’s writing development and manually crafted more advanced challenges, these systems use algorithms to create dynamic knowledge maps that continuously assess and adapt content. The underlying pedagogical insight remains the same: learning sequences should respond to demonstrated mastery rather than predetermined schedules.

As well, my adaptation of teaching different learning styles, cultural references and regional metaphors — shifting from desert imagery in Tucson to Le Guin’s misty forests in Portland — intuitively applied what adaptive platforms now formalize through multiple modality presentations. Where I observed student engagement and adjusted my metaphors accordingly, modern systems like Articulate Rise track interaction patterns to determine whether a learner better engages with visual, textual, or interactive content. However, these systems still can’t easily discern the cultural resonance of content — they might know a student prefers videos to text, but not that desert metaphors would feel alien to a Portland teenager.

As an English teacher, I developed an intuitive taxonomy of feedback — knowing when a struggling writer needed encouraging comments about ideas before addressing grammar, or when an advanced student was ready for stylistic coaching. Modern feedback systems like Grammarly now attempt to formalize this adaptivity, providing different levels of suggestions based on detected writing patterns. However, these systems still struggle with what I found most essential: connecting feedback to a student’s emotional relationship with writing. They can identify a grammatical error pattern but not recognize when a student’s bold metaphor represents a personal breakthrough worth celebrating before addressing technical flaws. Part off my teaching essays model was to always give students both a numeric score from the rubric but also to give them, as Clifford Geertz, the anthropologist, would call “deep and rich” feedback. I’d speak to the technical strengths and areas to work on, as well as speak to them on a human level about the content–supporting them if they took a particularly courageous emotional risk in their writing, talk about how something they wrote could fit into their future professional goals, or ask them personally reflective questions that I didn’t want an answer to but that their writing made me wonder about.

Every experienced teacher develops what might be called ‘predictive intuition’ — that sense that a student is struggling before formal assessments confirm it. I could often tell from subtle changes in engagement, question patterns, or even body language when a student was losing confidence in their writing abilities. Modern learning analytics platforms like D2L Brightspace now formalize this through data patterns, tracking engagement metrics, submission timing, and performance trends to predict challenges before they manifest in failing grades. What these systems gain in comprehensive data analysis, they lose in nuanced human observation — they might note that a student’s login frequency has decreased but miss the slight shift in writing tone that signals disengagement.

Developing my culinary world literature course required carefully selecting texts that would resonate with students’ interests while introducing new perspectives. I intuitively balanced familiar reference points with challenging new concepts. Adaptive platforms like Realizeit now formalize this curation process, assembling personalized learning paths by drawing from extensive content libraries. The system might determine that a learner responds better to case studies than theoretical explanations, automatically prioritizing practical examples. However, these systems still lack the cultural intuition to know that Gabriel García Márquez might resonate deeply with a student who has never articulated their connection to Latin American magical realism, a phenomenon that was all too common. So often my Latino students struggled deeply to understand academic concepts of metaphor, only to realize, once they did, that their cultural lives are deeply steeped in metaphor.

In my classroom, I intuitively created groupings that balanced students’ strengths and weaknesses — pairing strong analytical thinkers with creative storytellers to create writing partnerships that benefited both. Modern platforms like Packback now use algorithms to create these connections across digital environments, identifying complementary skill sets or knowledge areas among learners who might never interact otherwise. These systems can scale peer learning in ways a single classroom cannot, but they miss the human insight that sometimes guides effective grouping — like knowing when two seemingly incompatible personalities might challenge each other productively.

The truth of all of these parallels is that much of the human approach is based on a gut feeling. There are many highly effective and powerful decisions that I made with teaching that, especially early on, were based on my gut. In later years, when I’d reproduce those moments I’d be able to intellectually grasp why it works and formalize it. We are able to make intuitive leaps that are based on our embodied and lived experiences that AI cannot. However, on the flipside AI offers human educators tools that greatly enrich and enhance our teaching capabilities. I see the question of: “Is AI going to replace human workers?” to be inappropriately reductive in many industries, specifically when it comes to teaching and instructional design.

The Human Touch

My classroom approach always followed a careful progression: direct teaching for new concepts, guided whole-class discussions, collaborative small group work, and finally independent demonstration of mastery. This gradual release model offers instructional designers a sophisticated framework for balancing automation and human facilitation. Systems might automate the delivery of foundational content and basic practice, but human facilitators should guide the crucial transition from understanding to application. As learners progress toward independence, automation can handle routine assessment and feedback, while human intervention focuses on addressing conceptual misunderstandings and pushing advanced thinking. The key insight is that neither full automation nor complete human facilitation is ideal — rather, the balance should shift throughout the learning journey, with more human guidance during initial concept introduction and critical transition points, and more automation during practice and routine assessment.

However, this method did not always apply to the needs of every student. My most successful adaptations always began with absolute clarity about the essential skills students needed to master. For writing, regardless of how I adapted delivery, the fundamental objective of coherent, evidence-supported arguments remained constant. Similarly, instructional designers must establish clear, measurable learning objectives before introducing adaptive elements. Adaptation should vary the path, not the destination.

Even with my most independent learners, regular check-ins prevented them from going too far astray. I learned to schedule these moments strategically — before challenging concepts, after significant failures, and during transitions to new skills. Instructional designers should map the learner journey to identify similar critical moments where human intervention provides maximum value: onboarding to the system, transitions between major concepts, after assessment failures, and during application to real-world contexts.

My approach to teaching writing evolved constantly based on student results and feedback. When I noticed students consistently struggling with thesis development despite mastering structure, I created targeted interventions. Effective adaptive systems similarly need mechanisms to evaluate their own effectiveness and evolve. Designers should incorporate not just performance metrics but qualitative feedback channels that help the system and its human guides understand why certain adaptations succeed or fail.

I discovered that even the most perfectly tailored instruction fails when students feel no ownership over their learning journey. My most successful adaptations preserved student choice — whether selecting essay topics or choosing which skills to develop next. Instructional designers must balance algorithmic optimization with learner agency, perhaps allowing learners to choose between equally effective paths or determine their own pace within reasonable boundaries. Oftentimes these are the strategies that an educator is able to utilize to get through learning sticking points.

The Future of Adaptive Learning?

Throughout my teaching career, I witnessed educational technology evolve from basic computer labs to sophisticated online platforms. Actually, when I student taught, in 2004, my classroom had chalkboards and a record player to go with the filmstrips. The record motor was so old that every so often I had to give the record a push to keep it going. Each advance promised to transform education, yet the human element remained essential. As we look toward AI-enhanced adaptive learning, this pattern suggests both tremendous possibilities and important limitations.

As natural language processing and emotion recognition improve, AI will increasingly handle aspects of learning that once required human insight. Where I once had to manually assess a student’s writing voice to determine readiness for creative forms, future systems might analyze linguistic patterns and confidence markers to make similar judgments. However, the most profound moments in my classroom often came from authentic human connection — recognizing a student’s cultural background in their metaphors or connecting their writing to their personal struggles. Even advanced AI will struggle to replicate this authentic recognition.

Retrieval-Augmented Generation systems represent a particularly promising direction for adaptive learning. By combining large language models with specific knowledge bases about learning progressions and subject matter, these systems could create truly personalized learning pathways grounded in accurate information. Where I once had to rely on my memory of what worked with similar students in the past, RAG systems could analyze thousands of learning pathways to identify optimal approaches for each new learner while maintaining connection to verified content. As I get older it’s probably better that I’m not the only RAG in the system.

My experience teaching diverse populations heightened my awareness of how educational approaches can either reinforce or challenge existing inequities. For example, the very cultural adaptations that made my teaching effective — switching between desert and forest metaphors, incorporating culturally relevant literature — required awareness of how cultural background shapes learning. Algorithmic systems risk amplifying biases if they optimize for immediate performance rather than long-term development, potentially steering students from certain backgrounds toward lower-achievement paths. Instructional designers must build systems that recognize and address these ethical dimensions, perhaps by explicitly evaluating recommendations for bias before implementation.

I envision future adaptive systems that recognize the irreplaceable value of human connection while leveraging technology for what it does best. Rather than replacing teachers, these systems might free them from routine tasks to focus on the high-value human elements — building relationships, providing nuanced feedback, nurturing motivation, and modeling ethical thinking. The most effective systems will likely combine algorithmic adaptation for content delivery and basic assessment with regular, meaningful human interactions designed to inspire, challenge, and recognize each learner’s unique journey.

Conclusion: The Adaptive Designer

My journey from intuitive classroom adaptation to understanding systematic approaches parallels the evolution from traditional teaching to technology-enhanced learning design. The insights gained from twenty years of adapting to diverse students and contexts provide a foundation for approaching instructional design with both technological sophistication and human wisdom.

The metacognitive awareness I developed through years of teaching — constantly reflecting on what worked, what didn’t, and why — translates directly to effective instructional design. This reflective practice enabled me to move from intuitive adaptation to systematic approaches without losing sight of the human elements that make learning meaningful. As an instructional designer, I bring both systematic thinking and human insight to creating learning experiences that respond to individual needs.

The fundamental insight that drove my classroom adaptation — recognizing that different learners need different approaches to the same material — applies equally to workplace learning. Just as I created multiple entry points to literature for diverse students, instructional designers must recognize that organizational learners bring varied experiences, preferences, and needs to professional development. The principles of clear objectives, meaningful feedback, appropriate pacing, and cultural relevance transfer directly from classroom to corporate learning.

The most valuable lesson from my teaching experience is the importance of balancing structure with flexibility. As a teacher, I created systems that provided necessary guidance while allowing for personalization and discovery. As an instructional designer, this dual role continues — creating systematic learning frameworks while ensuring they remain responsive to human needs. The adaptive designer must be both architect, building systems that function effectively at scale, and guide, ensuring these systems serve genuine human development rather than mere efficiency.

From the desert metaphors of Tucson to the misty forests of Portland, my teaching journey required constant adaptation to new contexts while maintaining core principles of effective learning. Similarly, as learning technology evolves, the adaptive designer must continuously incorporate new tools and approaches while preserving the essential human elements that make learning transformative. Whatever technological advances emerge, the need for thoughtful adaptation — attuned to both individual learner needs and broader human development — will remain central to effective learning design.