How to Choose Your Wedge, Respect Context, and Build Products That Change Real Trajectories in EdTech
- Krzysztof Kosman
- Mar 13
- 6 min read
Updated: Apr 8
AI is transforming education at a time when systems are already stretched thin. We face teacher shortages, student disengagement, rising dropout rates, and an impending wave of job disruption.
In his conversation on the EdTech Dots Podcast, Greg Margas articulated a tension that many builders experience. We are overwhelmed by ideas and hype, yet we often lack focus on the few things that will truly matter in the next decade.
This article extracts several key themes from Greg’s insights and reframes them as a practical guide for founders, developers, and product leaders working at the intersection of software, AI, and education.
Watch the full EdTech Dots conversation with Greg Margas: https://youtu.be/M-86DgKOo7Y
Focus in an Overstimulated EdTech Market
The EdTech and AI space is filled with enticing buzzwords: gamification, neuro-edtech, adaptive learning, and copilots for everything. However, Greg pointed out that some of the most successful businesses in this space are built on narrow wedges. Examples include:
A single gamification plugin that enhances engagement in existing courses.
An accessibility plugin that makes e-learning usable for learners who were previously excluded.
These are not grand visions of “reinventing education.” They are focused solutions to specific problems.
What This Means for Builders
Pick One Problem: Choose one problem for one user in one context. For instance, “Help first-year CS students complete weekly exercises” is a better starting point than “fix computer science education.”
Anchor on Outcomes: Design around completion rates, concept mastery, and reduced manual teacher effort. Avoid focusing on metrics like the number of prompts or minutes spent in the app.
Ignore 90% of Trends: You don’t need to add a “community,” a marketplace, and a chatbot in version one. Focus is not just a product choice; it’s a cultural constraint on what you allow into your roadmap.
In an overstimulated market, clarity about the one problem you exist to solve is a competitive advantage.
Local-First, Global-Aware
Greg also emphasized a point that many globally ambitious teams often overlook: education problems do not transfer seamlessly between contexts.
A solution that works beautifully in one Polish university may fail entirely in a rural school in another country. This isn’t due to poor code, but rather because:
Curricula differ.
Schedules and assessment cultures vary.
Infrastructure and bandwidth are inconsistent.
Incentives for teachers and institutions differ.
What This Means for Builders
Pilot Deeply: Focus on a real department, a real school, and a real community. Avoid the abstract notion of a “global learner.”
Design with Constraints: Assume uneven bandwidth, mixed devices, and older hardware. Recognize that teachers may have very limited time and support.
Keep Solutions Configurable: Ensure your product can adapt to different calendars, grading scales, languages, and policies without needing a complete rebuild.
“Local-first, global-aware” is a more honest and scalable approach than “global from day one.” The goal is not to avoid specificity but to start specific while leaving room for generalization.
Hard Problems Are Layered Problems
One concrete challenge Greg discussed is university dropout rates. Institutions have struggled to reduce these rates for over a decade, achieving only partial success.
Understanding Dropout as a Multi-Layered Issue
Dropout is not just about learning materials or user interfaces. It involves financial stress, mental health, family obligations, poor academic preparation, and a lack of belonging—all of which interact.
In an AI-driven environment, it’s tempting to claim, “We reduce dropout with personalization and nudges.” However, this oversimplification erodes trust.
What This Means for Builders
Treat Problems as Ecosystems: View dropout as an ecosystem rather than a simple toggle. Your product may influence one aspect, such as early risk detection or smoother transitions between modules. Be honest about what you can address.
Be Evidence-Minded: Focus on logging meaningful signals, such as missed key activities or changes in participation. Surface these signals to humans who can act on them.
Keep Humans in the Loop: Ensure that advisors, tutors, teachers, or peers can see what the system flags and decide on actions. Avoid creating an opaque model that overrides human judgment.
Respecting complexity doesn’t mean you can’t contribute. It means being clear about which layer you’re addressing and designing your system to work with existing human infrastructure.
Equal Access as a 10-Year Benchmark
Greg proposed a simple ten-year test: in a decade, will a child in rural Poland have access to learning quality comparable to that of a student at an elite university? More broadly, will we have fewer “lost geniuses”—individuals whose potential was never realized due to their circumstances?
What This Means for Builders
Design for Imperfect Conditions: If your product only excels on modern hardware and fast internet, it likely excludes the very learners who would benefit most.
Think Beyond Institutional Budgets: Equal access is partly a business model choice. Consider tiers or sponsorship models that reach low-resource environments, allowing individuals to access value even if their institution does not purchase.
Be Intentional About Your Audience: There’s nothing wrong with building for well-funded schools, as long as you don’t claim to be “fixing access for everyone” while doing so.
Equal access doesn’t occur by accident. It happens when teams treat it as a design constraint and a guiding principle, rather than just a talking point.
AI Disruption and the Shift from Content to Pathways
Another of Greg’s insights challenges the narrative that “AI is just hype.” Job losses driven by automation and AI are real and already visible in some markets. Many estimates suggest that more than half the workforce will need significant upskilling or reskilling in the coming years.
If this is the case, then the role of AI-enhanced education is not merely to generate more content. It is to help individuals transition from one life path to another with dignity and confidence.
What This Means for Builders
Think in Terms of Pathways: A useful system doesn’t just answer today’s question; it helps individuals understand their current skills and gaps, identify a believable next role, and outline the steps needed to get there.
Consider Identity and Emotion: Losing a job is not just an information issue; it impacts identity and self-worth. Tools that assume a calm, confident learner may miss the mark.
Measure Success by Changed Trajectories: Did someone transition into a new role? Did they remain in a program that typically loses students? These metrics are harder to measure than “lessons completed,” but they are far more meaningful.
If AI in education results in more content but no clearer path, we will have squandered this opportunity.
Mindset, Fear, and Psychological Safety
Finally, Greg shared a personal story about internalizing the belief that he was “bad at maths.” This belief led to a freezing moment during a practice lesson in front of a class, making him hesitant to engage with teaching further.
This anecdote serves as a reminder that mindset and psychological safety are prerequisites for learning and for adopting new tools.
What This Means for Builders
Design Low-Stakes Practice Environments: Create spaces where students and teachers can try, fail, and try again without fear of public embarrassment.
Use Feedback to Build Confidence: Highlight progress and strategies, not just correctness. Celebrate small wins alongside addressing errors.
Treat Reluctance as a Signal: If teachers are slow to adopt your tool, the issue may not be usability alone. It could stem from fear of exposure, judgment, or replacement. Your product and messaging can either amplify or diminish that fear.
We cannot genuinely discuss “AI in classrooms” without acknowledging the emotional realities of the humans expected to use it.
Using This Moment Well
At the core of Greg Margas’s insights lies a simple, uncomfortable question: when we look back in ten years, will we be able to say we used this moment wisely?
For builders, this doesn’t mean “solve everything.” Instead, it translates into actionable steps:
Choose one real problem and one real context.
Understand its layers before promising to “fix” it.
Treat equal access and real human mobility as constraints, not slogans.
Build systems that create pathways for individuals whose lives are being reshaped by AI.
Design with the mindset and feelings of learners and teachers in mind.
If your next release helps even a small group of people move closer to these goals—be it one school, one program, one cohort, or one teacher—you’re on the right side of this moment.
Everything else is just noise.
