BridgeShift.AI

Inspiration

Our inspiration for this project came from the challenges we experienced during our own high school journeys. When students first enter high school, they are often overwhelmed by the amount of information they receive about college preparation and graduation requirements. Important details about volunteering and extracurricular opportunities can be difficult to find, confusing, or presented in a disorganized way.

We wanted to create a solution that makes these opportunities easier to access and understand. By providing clear and organized information, our project helps students spend less time stressing over logistics and more time exploring experiences that can support their personal growth and future goals.

What it does

BridgeShift AI acts as an intelligent translator for high school volunteer tracking. It allows school staff and counselors to take long, unformatted, and messy blocks of text from emails or flyers and instantly turn them into structured, plain-language opportunity posts.

For teachers, it provides a simple dashboard where they can paste raw announcement text, automatically get the important details pulled out, and verify the listings before students can see them. For students, it creates a personalized experience by matching opportunities directly to the interests and career categories they chose. Every recommendation explicitly tells the student why it was suggested, and the app dynamically shows direct action steps, like pre-written inquiry emails or one-click application links.

How we built it

We built our platform using a combination of low-code and AI tools to turn messy text into a secure database app. We designed the frontend user interface using Glide, which handles user sign-ins and automatically creates separate dashboards for teachers and students.

When a teacher pastes an announcement and clicks the organize button, Glide sends the text to our backend workflow on BuildShip. BuildShip runs a pipeline that sends the text straight to OpenAI GPT-5 with the temperature set to zero, which ensures the LLM acts as a reliable, accurate parser rather than a creative writer. The AI extracts nine core pieces of information against a strict data schema and saves it into our Airtable database with a Pending status so it stays hidden from students until a teacher approves it.

Challenges we ran into

Our biggest challenge during development was preventing the LLM from hallucinating or guessing details when a pasted announcement was missing information. In early tests, if a flyer forgot to list a contact email or a registration link, the AI would sometimes guess a realistic placeholder to fill the gap. In a high school setting, showing wrong dates or broken links would cause massive confusion and hurt the app's trustworthiness.

We solved this by giving the AI strict system instructions to never invent information, so if a detail is missing, it has to return Unknown. We then added logic to the Glide frontend so that if a field reads Unknown, the corresponding button like Apply Online or Email Organizer completely hides itself, keeping the page clean and accurate for students.

Accomplishments that we're proud of

We are incredibly proud of building a fully functional prototype that solves a genuine, everyday problem for our peers and school staff. Specifically, we are thrilled with our human-in-the-loop safety design. We successfully captured the speed of GPT-5 text processing but limited its authority, forcing every new opportunity to stay hidden as Pending until a real human teacher reviews the original source text and manually verifies the organization. We are also proud of building an explainable matching system that transparently tells students exactly why an opportunity is showing up on their dashboard, which builds a lot of user trust.

What we learned

This hackathon completely changed how we look at developing AI software. We learned that the true value of an AI tool isn't just about how advanced the technology is, but how intentionally you design your safety guardrails. We discovered how to effectively use LLMs as highly efficient assistants rather than completely autonomous decision-makers. By letting the AI handle most of the exhausting data entry and sorting, we can drastically reduce burnout for teachers while still keeping a human in complete control to ensure the platform stays safe for students.

What's next for M&A Innovations

Moving forward, M&A Innovations plans to expand BridgeShift AI past the hackathon phase. Our next immediate step is to build automated expiration workflows on the backend that read application deadlines and automatically archive past opportunities, completely eliminating manual cleanup for counselors. We also want to expand our categories to include critical community resources like local food banks and mental health support networks, transforming the app into an all-in-one plain-language support portal for high school communities.