DressLink

Inspiration

Dress for Success is an inspirational organisation that provides confidence, a safe space, and support services for women. However, behind the scenes, their team was spending hours manually processing referrals and client inquiries. Every email, form submission, and phone call required someone to read, interpret, and enter data by hand before a single volunteer could be matched to a client. We saw an opportunity to remove that friction entirely, so the people who work at Dress for Success can focus on what they do best: empowering women into the workforce.

What it does

DressLink is a platform connecting referral partners and Dress for Success staff. Partner organisations submit client referrals through a simple referral form. On the backend, AI automatically processes each submission, extracting key personal information, scoring urgency, assessing employment readiness, and recommending the right support services. Staff see a fully populated triage dashboard the moment a referral lands, with minimal manual data entry required.

How we built it

We built DressLink across four layers:

Staff & volunteer dashboard — A React site with Tailwind CSS and Vite, deployed to Azure Web Apps. Dress for Success staff and volunteers use this to view incoming referrals and their AI-generated triage profiles in real time.
Client intake form — A React + Vite app with a detailed intake form capturing demographics, employment background, and support needs. Used by both referral partners and Dress for Success directly to submit new clients into the system.
Backend & database — A Node.js/Express API connected to Azure Cosmos DB, providing a centralised store for all referral and client data.
Azure AI agent — An AI agent built on Azure that processes incoming client data, generates a full triage profile — urgency score, employment readiness, primary barriers, recommended services, and next steps — and writes the enriched record to the database, where it immediately surfaces in the dashboard.

Challenges we ran into

Our initial idea was to poll the Dress for Success Facebook page and pull interactions to generate new leads via a Logic App. We ran out of time before we could get it fully integrated, but it remains a strong avenue for the future.
We had many ideas early on and were unsure which direction to pursue. This was resolved once we spoke directly with Dress for Success, which helped us narrow our focus and align on the right solution.

Accomplishments that we're proud of

How well the team worked together, we divided tasks effectively and brought everything together by the end.
An end-to-end referral flow working in production, from partner form submission through to a fully AI-triaged record in the dashboard.
AI triage output that maps directly to the fields Dress for Success staff actually use, reducing onboarding friction for the organisation.
A clean, accessible dashboard that reflects the Dress for Success brand and feels approachable for the community organisations referring clients.
A solution that is inherently adaptable — the same pipeline can be extended to other intake channels such as social media, automated phone calls, and WhatsApp.
Growing our confidence with AI, what started as curiosity became hands-on capability, and by the end we were building with AI rather than being intimidated by it.

What we learned

We worked with technologies we hadn't been exposed to before, such as Azure, and gained valuable knowledge by learning from team members who had prior experience.
We learned not to let deadlines and challenges disrupt team momentum, instead of letting pressure fracture the workflow, we worked through obstacles together and reached our goals.
Planning is just as important as execution. The time we invested in scoping and aligning early on paid off throughout the build.

What's next for DressLink

Email ingestion - processing referrals sent directly via dress to success's email
Volunteer matching - using the triage profile to automatically suggest the best-fit volunteer for each client
Expanding Avenues - intergrating AI-agent into social media or automatic phone calls to take out more manual processing