RecomAid — Structured Recommendation Intelligence Platform
Confidential, structured, AI-assisted recommendations and decision-support—built on auditable inputs, templates, and role-based control.
Problem Statement
High-stakes recommendation and evaluation workflows are slow, inconsistent, and difficult to scale. Requests arrive unstructured; supporting evidence is scattered; writers rewrite from scratch; and institutions lack standardization, quality control, and confidentiality guarantees. Generic AI tools don't solve the core problem: structured evidence capture, controlled generation, and repeatable output quality within strict access boundaries.
Economic Impact
Cuts drafting time while increasing structural consistency and output quality; enables standardized templates and repeatable letter quality across users; reduces coordination overhead between requesters and writers through structured workflows and role-based access.
Core Capabilities
- Structured request intake (experiences, evidence, outcomes) with controlled visibility
- Role-based access control across Student / Professor / System Owner
- Template library + personalized writing styles and reusable snippets
- AI-assisted drafting with human-in-the-loop review and edits
- Auditability: grounded letter generation from captured inputs (not free-form guessing)
- Payments-ready workflow for paid requests (Stripe / LemonSqueezy compatible)
- Multi-section platform evolution (Letters, Applications, Knowledge) on shared infrastructure
Architecture Overview
Single application with shared backend infrastructure and clear product separation by routes/modules. React frontend + FastAPI backend with PostgreSQL (Supabase) persistence. A shared services layer handles auth, billing, templates, and AI orchestration (OpenAI configurable). Product areas (Letters / Applications / Knowledge) share infrastructure while isolating domain logic and tables where needed.
API
SaaS subscription (per-seat or per-organization) + optional usage-based billing
Integration Points
Ideal Client Profile
Education and credentialing organizations that generate or manage high-stakes written evaluations: universities, departments, scholarship programs, bootcamps, professional training providers, and career services teams—especially where confidentiality, standardization, and throughput matter.
Pilot Structure
2–4 week pilot: (1) select one program/department and define templates + rubric, (2) onboard a small set of professors/writers, (3) run real requests end-to-end with AI-assisted drafting + review, (4) measure time-to-delivery and quality consistency, (5) define rollout plan (SSO, governance, and scaled template library).