CASE STUDIES

Systems built for real operations, not pretty demos.

Each engagement solves a manual operational problem with an actual workflow, interface, and measurable shift in how the team works.

BLOCKCHAIN / CONTENT

Yap Market

3-Level AI Content Intelligence System

PROBLEM

Yap Market needed to evaluate thousands of pieces of user-generated content on a blockchain-based rewards platform. Manual review was impossible at scale, bot accounts were gaming the reward system, and quality scoring was inconsistent.

SYSTEM BUILT

Level 1 content quality scoring using OpenAI for originality, depth, readability, and relevance.
Level 2 bot detection layer that flags non-human submission patterns and suspicious activity.
Level 3 performance analytics dashboard for engagement, reward fairness, and creator quality trends.

OpenAI API, Python backend, custom dashboard

OUTCOME

3 moderation layers automated

Bot detection improved materially over manual review

Reward decisions became data-led instead of subjective

SHIPPING / LOGISTICS

Star Ship Management FZCO

Shipping Ops Automation V1

PROBLEM

Star Ship Management was manually tracking unpaid invoices across scattered email threads, missing inquiries in a busy inbox, and sending service updates through ad-hoc WhatsApp messages. All three workflows depended on memory.

SYSTEM BUILT

Weekly invoice follow-up scheduler that stops when invoices are marked paid.
Inbox inquiry tracker that logs inbound requests and flags unanswered inquiries after 48 hours.
Broadcast campaign builder with approval flow, send log, and performance visibility.
Invoice creator that generates branded PDFs in USD, AED, and INR.

Next.js, Supabase, n8n, React PDF, Gmail API (Phase 2)

OUTCOME

3 manual workflows consolidated into one panel

Zero missed inquiry tracking after automation

Campaigns moved from ad-hoc blasts to a governed process

MINING / RESEARCH

First Canadian Graphite

Competitor & Market Intelligence Bot

PROBLEM

The team needed recurring graphite market intelligence, competitor monitoring across multiple data sources, and clean internal reports. Analysts were losing days on repetitive desk research.

SYSTEM BUILT

Agentic research workflow using OpenAI with web search inputs and source synthesis.
Competitor monitor for filings, press releases, and public market signals.
Structured report generator covering market share, price movement, and competitor positioning.
Natural language query interface for on-demand research questions.