Class 10: September 17, 2026 · ENTR 5000
Notes from September 17, 2026. These notes record the discussion and instructions given that day. For current requirements and due dates, use the assignment guidelines and semester schedule.
Sprint 1 presentations: turning ideas into useful learning
Today’s presentations showed a range of early ventures, prototypes, and AI workflows. The recurring question was how to turn an interesting idea into a focused endeavor that can be tested with real people. The notes below combine the discussion with general instructor feedback on the presentations.
Start with a business problem and a reachable customer
A large vision becomes more useful when its first audience is specific and accessible. Personal experience and existing relationships can provide a starting point, but they do not establish that others share the problem or will use the proposed solution.
Distinguish a need from a nice improvement. People may agree that something sounds useful and still do nothing. The benefit must be meaningful enough to overcome the effort of changing their behavior—the “activation energy” between interest and action.
Research can change the question. If people describe a different frustration from the one you expected, revising the idea is learning. Explain what you looked for, what you found, and how that changed your next step. A proposed survey, a conversation expressing interest, and an observed product use provide different kinds of evidence.
Existing solutions and blockers are research opportunities
“Does this already exist?” is a starting question. Study what people currently use, including informal workarounds, and why a particular audience might choose something else. Existing competitors can leave openings in focus, access, trust, distribution, or product approach; their existence alone does not settle whether an opportunity is worthwhile.
A marketplace also faces a chicken-and-egg problem: buyers need useful supply, while sellers need buyers. The discussion suggested starting with an accessible community and learning from existing marketplaces. More broadly, ask how others have solved a similar blocker. AI can help generate and compare options, while real constraints and evidence guide the choice.
Scope something you can do, measure, and learn from
Several strong approaches reduced a large problem to one workable first step: a manual process, a narrow user group, a small assessment, or one interaction. The first version need not contain the eventual automation or every feature. Its purpose is to make a meaningful question testable.
Clarify what is working now, what is a prototype, and what is planned. A screen or demo can make a concept tangible without yet showing that it delivers customer value. When using charts, simulations, or test environments, explain what the evidence represents and the limits of the claim.
Use AI to support deliberate work
Useful progress included learning to save and deploy work, organizing research, specifying inputs and outputs, testing interactions, and connecting services. Naming a tool is less informative than explaining what you used it to accomplish and what you learned when it failed or required correction.
Keep ownership of the direction and assess the output. AI-generated slides can be clear, focused, and sound like the presenter when they are edited thoughtfully. Similarly, an automated workflow needs an appropriate scope, schedule, and check on useful results; recurring activity can consume a budget without advancing the project.
Scope the workflow before building it: what information is needed, what should happen, and what would count as success? When a data source or integration is unavailable, consider whether a smaller or manual version can answer the next question.
Instructor feedback across the presentations
Strengths to repeat: clear openings that connect the audience to a problem; focused, readable visuals; concrete comparisons; a specific first customer; honest descriptions of blockers; and adapting calmly when technology or the room interrupts the talk. Interactive materials can help explain a relationship when the interaction serves the point.
Opportunities: give more attention to what changed, what you learned, and what you will test next. Simplify specialized language so classmates can understand why the work matters. Keep ambitious possibilities separate from current results. Use AI to assess and revise a slide against the audience’s needs, then make the final editorial decision yourself.
Peer feedback is part of the learning. Ask a useful question, name a specific strength, or offer a suggestion that helps the presenter advance the work. You do not need to solve the whole venture to contribute.
The instructor is considering how presentation timing and the number of class meetings can leave more space for learning and next steps. No revised format was assigned today. Keeping shared materials and class summaries connected in Slack will help preserve useful examples and feedback.
Materials and next class
Student presentation materials were shared in #26f-5000-jw. There was no separate instructor slide deck for this AI class. For presenting from another computer, test your viewing link beforehand; a memorable short URL can make access easier. The services mentioned were TinyURL and Bitly.
Tuesday, September 22: students who did not present today will go first; keep your presentation ready. The class then moves into the next sprint. No new quiz, reading, or separate homework deadline was announced in this recording. Course communications run through Slack; grades and formal course submissions belong in eLC.