We chose the WIAL problem because it felt real in a way most hackathon prompts don’t. WIAL isn’t asking for a flashy app or a new AI gimmick, they already have a working organization, global chapters, and an LMS they’re happy with. The actual problem is harder: how do you maintain a consistent global presence while letting local chapters operate independently without everything becoming fragmented? That tension between global control and local autonomy is what pulled us in.
We’ve all seen what happens when organizations scale across regions, each chapter builds its own site, branding drifts, content quality varies, and it becomes harder to discover people or resources across the network. Instead of building something new for the sake of it, we focused on solving that structural problem. The goal was to create a system that gives chapters real flexibility, but within a governed framework that preserves WIAL’s identity.
We ended up building a governed global chapter platform, not just a website. That distinction guided most of our decisions. We intentionally avoided rebuilding the LMS or adding course delivery features, because that wasn’t the problem WIAL asked us to solve. Instead, we focused on a clean public-facing site, a chapter creation and publishing workflow, a global template that chapters can’t break, a coach directory that works across regions, and a dues/payment flow that feels operational rather than staged.
We were also careful about where we used AI. Rather than adding it everywhere, we focused on places where it actually solves something meaningful. One of those is cross-lingual coach discovery. Since WIAL operates globally, users shouldn’t have to search in the “right” language to find the right coach. We built a system where someone can search in Portuguese or French and still find relevant coaches whose profiles may be written in English, with clear reasoning for why those results match.
The second major AI piece is chapter content generation. Starting a chapter shouldn’t require writing an entire website from scratch, especially across different languages and regions. We built a flow where chapter leads can input basic information like location, audience, and focus, and get structured, culturally adapted content to start from. At the same time, we kept a strict boundary, nothing is published without human review. The goal wasn’t to replace ownership, but to reduce friction.
One of the biggest challenges was making the product feel real instead of like a collection of demo features. We spent a lot of time simplifying the UI, cleaning up content, and making sure different parts of the system actually connect. The governance model was also harder than expected, it’s easy to say “global template with local edits,” but much harder to decide what should be locked, what should be editable, and how content moves from draft to published in a way that feels believable.
Another challenge was AI reliability in a live setting. The cross-lingual search is one of the strongest parts of the project, but also the most fragile if it doesn’t behave consistently. We had to make sure results were stable, explanations made sense, and the experience didn’t fall apart with slightly different inputs, especially since this is something we demonstrate live.
What we learned most from this project is that good solutions aren’t about adding more features, they’re about respecting the constraints of the problem. We learned to build around existing systems instead of replacing them, to prioritize structure and governance over feature count, and to use AI where it adds clarity rather than noise. We also took accessibility more seriously, integrating multilingual support and text-to-speech so the platform is usable beyond a narrow audience.
At the end of this, what we built isn’t just a nicer website. It’s a system where WIAL can maintain a consistent global presence, chapters can move faster without starting from scratch, and users can discover coaches and resources regardless of language or location. That combination of governance, accessibility, and discovery is what makes this feel like something that could actually be used beyond the hackathon.
Built With
- elevenlabs
- next.js
- openai-(gpt-4o
- propelauth
- react
- stripe
- text-embedding-3-small)
- typescript
Log in or sign up for Devpost to join the conversation.