By Jefferson Pestronk, Executive Director, Modern States Education Alliance

A few years ago, I read a blog post by Steph Smith, growth marketer at Nvidia, titled “How to Be Great? Just Be Good, Repeatably.” In the blog, Smith writes:

… I realized that it was not the sporadic highs that were exceptional, but instead the long hauls; the sequences of events that seemed minimal at each juncture, but compounded into major gains. This led me to think further about what greatness truly means. I’ve come to learn that it’s not about overnight successes or flashes of excellence, but periods of repeatable habits … Perhaps “great” is just “good,” but repeatable.

This idea has remained with me, and the concept is particularly relevant to the mission of Modern States: we think great learning requires regular and productive engagement, and our organization exists to give every motivated learner the chance at that engagement without the cost of and friction within the higher education system. We ourselves are incessantly focused on making targeted improvements to our courses and the support we offer to students as our path to great courses.

As we shared in our year-end post, 2025 was Modern States’ best year ever. More learners, more CLEP exam passes, more college savings. Those outcomes were the result not of a giant change to what we do, but of multiple smaller tweaks: aligning the design of our courses to create a consistent learning experience, shifting the way our quizzes and final exams work to encourage mastery, simplifying our voucher request process cut the turnaround time in half, and more. Not one great change, but lots of good changes we can do over and over.

What does this mean for 2026? And why was I thinking about this idea in particular?

One of the themes of 2026, everywhere one looks, is that AI is going to change everything. AI is exactly the kind of “big bet” moment that tempts organizations to chase the transformational thing and to neglect the impact that can come from repeatable systems. The simpler principle that greatness is actually just good, consistently, becomes especially important as we navigate these changes.

I have gravitated towards a piece by author Rick Hess, with the typically provocative headline “AI Changes NOTHING About What Students Need to Learn.” Hess argues that knowing things and building traditional academic skills will remain as important as ever in forming foundations for individuals to think about and engage in the world.

Taking an example from the past: using a calculator might allow someone to avoid learning the math the calculator performs, but someone who does learn the math will remain much better equipped to use the calculator effectively. Just as learning algebraic principles helps students recognize when their calculator gives an impossible answer, mastering historical analysis and drawing on a base of actual knowledge helps students recognize when AI-generated products miss the mark. AI can draft an answer, but knowledge is what lets you judge whether it’s coherent, accurate, and appropriate.

The skepticism in this perspective does not mean we’re shying away from AI at Modern States. But it shapes how we’re thinking about it in 2026.

We remain focused on building outstanding and accessible courses in traditional academic disciplines. Having access to college math, science, writing, and history is as critical as ever; while there is concern about the value of a college degree, most research continues to show a large earnings premium and future projections suggest it will endure.

In this context, AI becomes a tool to help us build and to help learners learn. In 2025, we started using AI tools to accelerate the efforts of the college faculty who help us build our courses, particularly in areas like developing large pools of assessment questions. We’re leaning into this type of use: humans driving and overseeing the process but AI accelerating the pace beyond that of any individual, and we’re using it to shape our approach to annual course enhancements that are responsive to learner data and feedback.

We also plan to pilot a learner-facing AI tutor (Lumi Tutor) in a handful of our courses. Connected to our course content, an AI tutor can help learners get unstuck by offering a different explanation of a challenging concept or a connection back to prior learning. It can offer in-the-moment assessment to support regular practice, along with explanations of where learners went wrong if they miss questions. We’re not chasing AI magic – we’re seeking fewer stuck moments, more relevant explanations, and more practice opportunities at our scale of hundreds of thousands of learners, repeatably and affordably.

All of this will be grounded in intentional data collection and evaluation. We have a theory of why we’re making specific changes, but we recognize that not all changes result in improvement. Starting small with changes, testing to see if the changes improve outcomes, and then expanding the ones that do disciplines our effort. In addition to our internal monitoring, we plan to launch formal research and evaluation partnerships this year that will expand our capacity to learn as an organization and add value to the broader field.

This captures only a fraction of what Modern States will be up to in 2026. We plan to revise and improve our implementation materials for educators and schools. We also have new partnerships we’re exploring, allowing us to bring Modern States to specific groups of learners and targeted geographies.

This year ahead of us is about strengthening and extending the systems that help learners earn college credit, save money, and believe in their ability to succeed. As Modern States continues to grow, so does our responsibility to be good – repeatably. It’s the best way to achieve greatness, and the best way to make 2026 our most successful year yet.