How AI and Vibe Coding Transform Product Management
By Adrian C. Ott
Product management (PM) is being reshaped by AI at a pace that’s breathtaking. The speed from idea to prototype has fundamentally changed how our students learn and practice their craft.
Many of our 麻豆村 Integrated Innovation Institute (iii) graduate students are AI Natives. With OpenAI premiering in late 2022, these students already used AI in their studies and some in their industry jobs. This raised an important question: how do we teach them to apply AI in product management when they’re already fluent in the technology?
Math didn’t stop being taught once calculators were invented. Similarly, AI makes product development faster, but experienced PMs still need to know when an AI model hallucinates and suggests something unrealistic. In addition, Vibe Coding—the ability to develop software prototypes using AI prompts—was only coined in early 2025, yet it’s already becoming integral to future PMs’ ability to create more realistic product requirements.
This was our challenge.
Where AI Enhances Product Management
Throughout our seven-week course, students explored how AI could enhance key areas of product management:
Customer Discovery: Testing customer ideas, refining discussion guides, understanding customer environments and process flows, and asking better interview questions than simply “describe what you do.” The benefit? Deeper customer understanding faster, enabling interviews to focus on higher-impact questions.
Data Analysis: Identifying patterns in qualitative data from customer interviews, reducing bias, and surfacing opportunities that might not be readily apparent.
Vibe-Coded Prototypes: Better visualization of product concepts. In the past, we asked for basic mock-ups on paper or simple wireframes. Now it’s possible to easily visualize ideas and create a stronger foundation for communication with software engineering teams.
AI as the Product: Guest speakers discussed how the PM approach differs when an AI model is the core product vs. AI as an addition to traditional software or atom based product such as hardware.
Presentations and Stakeholder Buy-in: Faster development and refinement of ideas.
Competitive and Market Monitoring: AI Agents to monitor competitor actions and test demand for new functionality.
Course Structure and Student Projects
Our seven-week course followed a consistent pattern: each week began with a lecture on a PM topic, followed by a tutorial applying AI to that topic, then students applied those techniques to their team projects.
For their group projects, students selected a popular software brand. They applied course exercises to develop a new product concept that addressed significant customer pain points discovered through their research while aligning with the company’s Northstar objective.
Students selected companies including LinkedIn*, TripAdvisor*, Spotify*, Duolingo*, and YouTube Learning*. Their concepts ranged from applying AI to better match graduating students with job opportunities on LinkedIn, to integrated AI trip planning for small casual groups on TripAdvisor.
Industry speakers joined us every week, sharing practical techniques and trends beyond the classroom related to that week’s PM topic.
Innovating the Classroom: AI and Vibe Coding Techniques
We incorporated AI into the PM role through several innovative approaches:
AI Bibliography
We encouraged students to test and apply AI for every assignment, knowing they’d need this for their future PM roles. Rather than prescribe specific tools, we encouraged experimentation and asked students to document their learning: Which tools did they try? What prompts worked or didn’t work? Which tool provided better results and why?
One student with deep AI agent experience noted that documenting and evaluating their approach in an AI bibliography was sometimes harder than performing the PM task the traditional way because “you had to pay attention to the tools and methods as much as the outcomes.” She found this approach beneficial.
Vibe-Coded Prototypes Required
We now require vibe-coded prototypes rather than basic wireframes or sketches—reflecting what the workplace will demand.
While I wasn’t surprised that students developed their prototypes rapidly, what impressed me was how the prototypes brought concepts to life in more realistic, sophisticated presentations for management and stakeholder evaluation and justification. It also allowed for faster, more iterative evolution of the concept.
AI Product Management Course Advisors
We recruited second-year students to voluntarily advise first-year students in our course on developing their product concepts and vibe-coded prototypes based on their AI experience. Each PM advisor was assigned to a project team. They also provided valuable insights to the teaching team on student readouts, needs, and helpful materials to share.
Second-year advisors gained valuable management experience and a resume differentiator for job interviews. First-year students valued hearing advice from their second-year peers, most of whom had PM and AI experience from their summer internships and previous jobs.
Because this course was held during mini 1 of their first year, the program also helped build community on campus by connecting students outside their first-year cohort. A big thank you and congratulations to our recent graduates who served as PM Course Advisors: Olivia Xiao, Nivashini Shivakumar, Kshitij Kakade, Keerthi Raghavendra, and Mustafa Saifee.
AI Tutorials for PM Tasks
Yaras Sheik, a graduating part-time iii student who also works in industry as a product manager, created three in-depth classroom video tutorials demonstrating how PM tasks could be enhanced with AI. This was helpful content not just for this course but will be useful for future PM courses.
Students Teaching Students: The PM AI Showcase Library
Like Yaras’s AI video tutorials, we asked students to create their own short tutorials using an AI tool of their choice for a PM task. This extra credit assignment resulted in most of the class creating videos to share with future PM students in the 麻豆村 iii program. These videos will also be helpful tools for showcasing their skills in future PM job interviews.
More than 17 tools were showcased and included software demonstrations using student accessible versions such as Loveable*, Cursor*, Bolt*, NotebookLM*, Miro AI*, Manus AI* and Claude*. The tasks covered the full spectrum of PM work. Here are a few examples from our class PM AI showcase:
Jessie Xiong used Claude and Lovable to create a landing page fake door demand test*.
Saloni Parekh applied AI to Slack to help product teams identify and address critical software issues more quickly*.
Harsha Gururaj demonstrated how Julius AI can analyze qualitative interview transcripts from customer interviews*.
Final Thoughts
Will AI eliminate PMs? Not anytime soon. While AI accelerates many tasks, one key learning emerged: humans must remain in the loop at critical junctures to guide AI and ask the right questions. In addition, not all products are 100% digital - traditional PM processes still apply even if there is software attached. Moreover, some software operates in highly regulated industries such as healthcare or medical devices in which software change processes need to be closely monitored to avoid adverse outcomes.
No matter the product, PMs must still garner support, alignment, and funding for their ideas—tasks requiring creativity and stakeholder engagement that AI cannot replicate alone.
AI and vibe coding tools free PMs from routine prototype building, allowing them to focus on what matters most: creative thinking and building stakeholder buy-in for their ideas. I am grateful to all the students, speakers, TAs, and advisors who participated in this course to bring everyone’s PM AI learning to a higher level.
*These are student projects created at 麻豆村 for educational purposes only. Carnegie Mellon and the work the students performed are not affiliated, endorsed by, or associated with the companies named in this article.