AI vibe-coding for product teams: Jorge Alcantara (AI Product Engineering, Zentrik)

In this hands-on session, Jorge Alcantara (AI Product Engineering at Zentrik) makes the case for why product managers are best positioned to lead the AI transformation in product development by mastering the art of vibe-coding.

Rather than replacing traditional specs, AI tools like Lovable and Magic Patterns help translate natural language into working prototypes — compressing weeks of effort into hours. Jorge walks through how to write effective prompts, structure requirements like a product brief, and build shareable, testable prototypes that accelerate team alignment and discovery.

This session is designed for product managers who are tired of being “Jira janitors” and want to get closer to customers, faster. 

Chapters
00:00 - Jorge Alcantara opens the session: why this is hands-on  
01:40 - The real bottleneck: PMs stuck in busywork  
03:00 - Why PMs are uniquely positioned to lead with AI tools  
04:20 - What “vibe-coding” (B-coding) actually means  
07:00 - From idea to prototype: visuals over PRDs  
11:00 - Writing prompts like a product manager  
14:00 - Jorge demos building a prototype using Lovable  
22:00 - What works, what breaks: lessons from practice  
27:00 - Best practices: iteration, screenshots, versioning  
33:00 - Prompting tips: using the right language and style  
36:30 - Publishing prototypes: SEO, domains, live links  
43:00 - Jorge’s final thoughts: using AI tools professionally

Key takeaways
— The core value of AI–assisted “vibe‑coding” tools isn’t to replace traditional specifications (e.g. PRDs), but to operationalise them — helping PMs translate ideas into live prototypes faster.
— PMs are uniquely positioned to lead this shift due to their command of natural language, stakeholder context, and user needs.
— Prompts should be treated like product requirements — explicit about who (persona), why (user problem), what (flow), and how (mock data, interactions, etc.).
— Use iteration, version control, and remixing: treat each prototype as a working draft, not final code.
— Visual input matters: share screenshots, mockups, and style references — the AI only “sees” what you give it.
— Keep scope realistic: AI‑driven prototyping excels for MVPs and internal tools, but isn’t yet fit for complex legacy codebases.
— Treat AI as a professional, collaborative tool — best used by PMs, designers and engineers together to accelerate discovery and alignment.