VSAC is a nonprofit offering low-interest loan programs as an alternative to high-interest private lenders. The students it serves face award letters that are nearly impossible to compare: one school may show a lower balance while assuming $30,000 in loans, another a higher cost with only $5,000. VSAC wanted students to photograph their letters and get instant AI analysis on their phones. The same platform would generate qualified loan leads, power a partner lead marketplace, and build a financial aid research database. So VSAC partnered with NineTwoThree to build Award Advisor.
Custom ML model combining image preprocessing, OCR, and LLM refinement to extract and standardize award letter data.
React Native app deployed to iOS and Android from one codebase, cutting development time and maintenance costs.
Tiered permission system giving VSAC admins full control and partners curated access to qualified leads.
Qualified lead generation showing partner organizations each student's exact financial gap for targeted loan offers.
Building a reliable AI tool for unstructured financial data on a nonprofit budget required solving problems no off-the-shelf solution could handle.
Award letters lack any standardization. Formats, layouts, and data presentation vary wildly, even within the same institution. Out-of-the-box LLM solutions could hallucinate numbers and rack up unsustainable costs on a nonprofit budget. VSAC needed a custom approach to extract accurate data from inconsistent documents.
Students needed mobile-first comparison tools, VSAC admins needed full platform control and analytics, and partner loan organizations needed curated access to qualified leads. Each audience required a different interface and permission level, all served from a single system.
The solution needed native-quality iOS and Android apps plus an intuitive desktop admin portal, all without building everything twice. Loan advertisements also needed hierarchical display, always showing nonprofit providers in the student's state first.
We assembled an expert team and ran collaborative workshops to validate requirements before writing any code.
We built a custom model combining image preprocessing, OCR, and targeted LLM refinement, avoiding expensive, hallucination-prone AI-only approaches.
We built databases mapping term variations to standardized categories, enabling accurate cross-school comparisons despite inconsistent award letter formats.
We developed once and deployed to both iOS and Android from a single codebase, cutting development time and maintenance costs while delivering native-quality performance.
We created distinct permission levels for VSAC admins versus partner organizations, with a lead marketplace showing each partner the exact financial gaps of interested students.
With the creation of their lead marketplace, VSAC can generate income from a source that was previously unavailable, allowing the project to be ROI-positive.
Students can instantly compare true costs, where they would have previously spent hours manually reviewing award letters and trying to reconcile inconsistent formats.
Award Advisor enabled the organization to review thousands more grant applications annually, significantly expanding their team's reach and impact across Vermont.
Award Advisor expanded VSAC beyond a single service into a platform operator, connecting students, VSAC admins, and partner loan organizations in one system.
.svg)
NineTwoThree AI Studio delivered the software ahead of the deadline. However, they have yet to see results. The team stayed on top of everything to keep the project on track and communicated via email, messaging apps, and online meetings. Moreover, every aspect of their work was of high quality.
Build a new platform, or add AI capabilities to existing software, we deliver solutions that work in real operational environments.
Meet with founders Andrew Amann & Pavel Kirillov
Custom AI implementation roadmap
No sales pressure
