Sam's take
I want to start with something Saul Nurick said to me a while back that feels more relevant than ever:
"You have to earn the right to automate."
I think he's right, and I think it cuts against where a lot of the assessment conversation is heading.
The current logic goes something like this: AI can do the rote stuff, so let's stop testing the rote stuff and push students up the ladder to critical thinking. Sounds reasonable. It's also, I suspect, wrong in the specific way that matters most for finance and real estate.
You cannot think critically about a DCF if you don't know, in your bones, why the denominator behaves the way it does. You cannot reason about a cap rate if you've outsourced the underlying arithmetic before you ever felt the relationship between NOI and value. Critical thinking isn't a separate skill that floats above content. It's what happens when you've internalized enough structure to notice when something is off.
So when I hear "let them use AI on the exam, we'll assess the higher-order stuff," my honest reaction is: with what? Higher-order thinking on top of what foundation? The students who get the most out of AI tools in industry are the ones who can smell a wrong answer. That nose comes from having done the work yourself, badly, many times, before automating it.
I could be wrong about this. It's possible the next generation builds intuition in some new way I can't yet see. But until I'm convinced otherwise, I think our job is to protect the part of learning where students earn the right to automate.
Teaching notes
Most finance and real estate courses are built on an assumption that doesn't hold up against the memory research: cover a topic in one chapter, run the problem set, move on, then revise everything the week before the exam.
A meta-analysis published this year by Murray, Horner and Göbel, looking specifically at mathematics learning, is the latest reminder that this isn't how retention works.
Two findings stand out:
→ Spreading practice on each topic across multiple sessions produces better retention than covering it all in one block.
→ Mixing different topics and skills together (interleaving) adds further benefit on top of that.
The uncomfortable implication: the standard textbook structure (one chapter, one problem set, move on) may be one of the least effective ways to help students retain what they're learning. Which is why so many of us spend the first two weeks of every semester reteaching time value of money to students who passed an exam on it four months ago.
Why this matters for how you spend your prep time: Most instructors I talk to are not short on content. They have decks, problem sets, worked examples, old exams, a textbook they like well enough. What they're short on is time to redesign the architecture of the course so practice is distributed the way the research suggests. Rebuilding a syllabus so week 9's quiz revisits week 3's material, while interleaving week 7's, is the kind of work that sounds simple and then eats a Saturday.
A few practical moves that don't require burning down your existing course:
Stop letting weekly quizzes be weekly. If your week 6 quiz only covers week 6 material, you've told students exactly when they can forget week 3. The smallest viable change: make every quiz 70% current week, 30% earlier material. Students complain for about two weeks. Then their exam scores improve, and they stop complaining.
Build a running problem set instead of a chapter-end one. In a corporate finance course, the WACC problem set in week 8 also includes two TVM problems and one bond pricing problem. You're not adding content. You're redistributing the practice you were already assigning.
Interleave within topic, too. If you're teaching DCF, don't do five DCF problems in a row. Do a DCF, then a comparable companies problem, then another DCF, then a precedent transactions problem. Students hate this in the moment because it feels harder. That difficulty is the learning. Bjork's "desirable difficulties" research has been telling us this for thirty years and we keep ignoring it because block practice feels productive.
Treat the cumulative final less as a checkpoint and more as the endpoint of a system. If students have been encountering week 3 material in weeks 5, 8, and 11, the final isn't a cram event. It's a confirmation.
The genuinely hard part, and the reason most instructors don't do this even when they know the research, is logistics. Writing the spaced quiz questions. Tagging which concepts appeared when. Making sure the interleaving is actually interleaved and not randomly clustered. Keeping it consistent across semesters when you tweak the syllabus.
We built Blended Teaching to handle this. The Blended AI assistant lets you assemble practice that's spaced and interleaved across the semester without you maintaining a spreadsheet. You design the course once. The platform handles the distribution.
If you teach a subject where retention still matters six months after the exam (and in finance and real estate, it almost always does, because the next course assumes it), the design work is worth doing. Whether you do it with us or with index cards.
Worth a look
Microcredentials are no longer a "trend" — they're the hiring signal
90% of employers prefer candidates with a microcredential over those without; 92% prefer candidates with a GenAI microcredential specifically.
CFA Institute updates the curriculum — AI, blockchain, ESG now in the canon
It might be time to start thinking about incorporating AI, blockchain, and sustainability into your curriculum.
Blended Teaching launches a new Financial Modeling digital textbook
A complete video-first approach to learning financial modeling.
Sam