AI-first learning · built around real university courses
Learn finance and economics the way you'll practice them.
Kumu pairs your professor's actual course materials with interactive labs that compute like the real Excel models, and with stage-by-stage guides for project work. The point is the journey, not the destination: the judgment you build getting there.
The doer–reviewer dilemma
Until recently, the first job in finance, marketing, consulting or accounting was an apprenticeship with a paycheck. Analysts built the spreadsheets, drafted the slide decks, checked the footnotes and wrote the first draft of the memo. Senior people reviewed that work, and the review was where the learning happened. Build enough models and you start to see the error before you can say why.
AI changes the bottom of that ladder. A model that took an analyst a weekend can now be drafted in minutes. Companies still need judgment, arguably more of it, because someone has to catch the errors in the draft AI produced, and those errors are easier to miss. Judgment, though, has always been built by doing. So the dilemma is this: the work that trained reviewers is the work AI now does first, faster, and with errors that are easier to miss.
Kumu's cases are built around that problem. You do enough of the work yourself to understand the model, then direct, test and revise what AI produces. Each case moves through the same five steps: frame, specify, build with AI, validate, and decide.
The order matters. The spec comes first so there is a standard to test the model against, and it is committed before the build so the order is on the record. All of it happens in the open on GitHub, so what you leave with is evidence of judgment: the spec, the audit trail and the decision, alongside the finished file.
AI can build the model. The judgment is yours to own.
Read the full essay: The doer–reviewer dilemma.
What's different here
The tutor reviews; it does not write
The tutor on every page knows which stage you're on. It explains concepts fully, critiques your reasoning, and walks worked examples — but it won't write your graded memo, spec, or analysis. That's the same line your syllabus draws.
Compute like the real model
Each lab runs the same math as the Excel workbook it accompanies — same conventions, same validation checks, same answers. Drag a slider until the intuition lands, then open the spreadsheet already knowing what every formula should produce.
Built around graded work
Each stage page carries the deliverable, a submission checklist that remembers your progress, and a short quiz. Weights and due dates are in your course schedule.
Subjects
Kumu is organized by subject, not course code — the same material serves any school's offering. If your syllabus sent you here, it linked the subject you need; if you arrived on your own, pick by what you are studying.
Student Onboarding
Before any course: set up GitHub, meet your AI tutor, tell it how you want to work, and start the professional portfolio — bio, resume, and the AI toolkit you'll carry into your career.
Micro & Macro Economics
One graded case-study engagement, two optional ones, and a team case study, tracing one arc — a price taker with no power over price, a patent holder who sets it, and a market where the profit turns out to be somebody else's rent. Plus live supply-and-demand and farm-profit labs.
International Corporate Finance
The Performance Ratios project: 9 stages from standing up a public portfolio repo to valuing a company you choose with a DCF skill you adapted and audited yourself.
International Finance & Securities
The FX Hedging project: frame a multi-million euro exposure, spec the hedge model, have an AI build it, load live market data, and validate the whole chain before making the CFO call — 6 stages, Design → Build → Validate.
International Economics & Trade
The import-a-product team project: source it on Alibaba, price it against Amazon, brand it with a real design system, survive the tariff, and sell it to your classmates. Two milestones, AI + GitHub throughout.
More Finance & Econ subjects
Financial statements & ratios for undergraduates, and beyond — the plan is a comprehensive Finance & Economics tutorial library, subject by subject, then topic by topic, at any school.
Looking for your course? Your schedule lists it by code: the course materials repository.
Using AI here (and its limits)
Honest by design
An AI model can be wrong, especially about specific numbers. Treat every AI output as a draft
to verify against the source, and log meaningful AI sessions in your prompt-log.md.
During this pilot the tutor runs in bring-your-own-key mode: click the ✳ button, open ⚙ settings, and paste an Anthropic API key (stored only in your browser). In the full product this moves server-side and disappears entirely.
Kumu has no in-page chat. The stage guides, labs, quizzes, and checklists all work on their own. Use your own AI (Claude, ChatGPT) as a coach, not a copier, and log meaningful sessions in your prompt log.