InquiryOps AI Workflow System
An independent, AI-assisted local portfolio prototype built around a technical-vocational school inquiry workflow. The system connects a responsive information site to persistent inquiries, private booking links, role-protected staff operations, consent history, grounded FAQ retrieval, and durable follow-up jobs. The AI boundary is intentionally narrow: without an API key, the assistant returns reviewed knowledge-base content; with the optional model adapter, the model may select only an approved answer ID. It cannot invent school facts, decide eligibility, or receive unnecessary personal details. The prototype uses synthetic records and does not claim client approval, live enrollment, production deployment, or business results.
Case Study
The challenge
Turn repeated school inquiries into an organized, privacy-aware workflow without letting AI invent admissions information or making a local prototype look like a live client deployment.
Approach
- Mapped the path from program discovery through inquiry, booking, staff action, follow-up, and opt-out
- Kept AI inside an approved knowledge boundary with a useful no-key fallback
- Separated domain logic from orchestration so the core system remains testable without n8n
- Added validation, authorization, consent controls, idempotency, retries, and human review paths
Evidence
- 51 domain tests and 13 end-to-end browser tests passed
- Seven n8n workflows imported and executed successfully in an isolated local runtime
- Architecture, requirements traceability, defect log, release checklist, and provider contracts
- Responsive evidence across 375, 768, and 1440 pixel viewports
Outcome
A working, production-shaped local demonstration that proves system design, AI boundary decisions, workflow automation, operational tooling, and QA discipline while keeping outbound delivery disabled and the project status explicit.