Deposify

Deposify – AI-Powered Tenancy Deposit Platform (Fintech/Regtech)

Ireland’s rental market lacks modern deposit management technology. The RTB (Residential Tenancies Board) dispute process takes 17+ weeks to resolve. No competitor combines AI-powered resolution with evidence validation. March 2026 regulations introducing 6-year minimum tenancies are transforming the market, creating both urgency and opportunity.

The Challenge

Design and build a complete AI-powered tenancy deposit platform that:

• Scales from MVP to enterprise (property managers, Approved Housing Bodies)
• Creates a distinctive, trust-building fintech aesthetic (avoiding generic “AI slop”)
• Handles complex dual-user journeys (tenants AND landlords)
• Integrates AI-assisted dispute resolution while maintaining GDPR Article 22 compliance (humans must accept AI recommendations)
• Meets RTB regulatory requirements for evidence handling and timelines

Task

Design and build a complete AI-powered tenancy deposit platform that:

1. Creates a distinctive, trust-building fintech aesthetic (avoiding generic “AI slop”)
2. Handles complex dual-user journeys (tenants AND landlords)
3. Integrates AI-assisted dispute resolution while maintaining GDPR Article 22 compliance (humans must accept AI recommendations)
4. Meets RTB regulatory requirements for evidence handling and timelines
5. Scales from MVP to enterprise (property managers, Approved Housing Bodies)

Tenant Screens

Landlord Screens

Key Actions

1. Design System Creation — “The Hex Theme”

• Developed a dark-mode fintech aesthetic with rose pink (#f5c0c0) accents
• Established JetBrains Mono for all financial data (deposits, IDs, timestamps)
• Created semantic status colors: Teal (success), Gold (warning), Coral (destructive)
• Built a complete component library with CVA (Class Variance Authority) patterns
• Ensured WCAG 2.1 AA accessibility compliance throughout

2. Core UX — Timeline-Based Milestone Tracking

• Designed the 4-Point Anchor Framework (Deposit Secured → Tenancy Start → Tenancy End → Resolution Deadline) with a planned Mid-Tenancy Checkpoint for the new 6-year minimum tenancies — addressing the evidence-freshness problem unique to long-term deposits.
• Created horizontal timeline components showing deposit lifecycle progress
• Differentiated tenant view (4 milestones) vs landlord view (6 milestones) based on complexity

3. AI Workflow Design

• Architected “AI-Assisted Resolution” flows (NOT “AI Arbitration” — terminology matters for legal compliance)
• Designed dual-party acceptance UI ensuring humans make final decisions
• Built escalation paths to RTB for disputed cases
• Created confidence scoring and explainability displays for AI recommendations

4. Technical Implementation

• Stack: Next.js 16, React 19, TypeScript, Supabase/PostgreSQL, Prisma ORM, Stripe Connect, AI SDK
• Used AI Agile framework and methodology with specialized agents (Sally for UX, Winston for architecture)
• Implemented Cursor + Claude (mcp & skills) for AI-assisted development workflows
• Built evidence handling with EXIF validation and SHA-256 hashing for tamper-proof records

5. Multi-User Dashboard Design

• Tenant Dashboard: Balance tracking, deposit cards, timeline progress, quick actions
• Landlord Dashboard: Portfolio management, property cards, tenant inbox, dispute alerts, Stripe Connect onboarding
• Messaging System: Threaded conversations with evidence attachments and AI resolution recommendations

LOGIN DETAILS

• [email protected] / landlord123
• [email protected] / tenant123

Results

Delivered

• Complete Hex Theme design system with documented tokens and patterns
• Timeline-based milestone tracking UX (a first in the deposit management space)
• AI-Assisted Resolution workflows compliant with GDPR Article 22
• Dual-dashboard system serving both tenants and landlords
• WCAG 2.1 AA accessibility compliance
• RTB-ready evidence export package generation

Market Position

• Zero AI competitors in Irish deposit management market
• First platform combining AI resolution + EXIF evidence validation
• Positioned for March 2026 regulatory changes (6-year tenancies increase deposit value 6x)

Technical Achievement

• Reduced typical 17-week RTB process to days via self-resolution
• Built complete component library with CVA variants for consistent styling
• Established patterns for handling 15 deposit states across database, UI, and timeline views

Key Learnings

1. Terminology matters in regulated spaces — “AI-Assisted Resolution” vs “AI Arbitration” has legal implications
2. Design systems prevent AI slop — Pre-defined tokens stop generic aesthetics from creeping in
3. Dual-user products need distinct but cohesive experiences — Same theme, different information density
4. AI-assisted development accelerates iteration — BMAD methodology with specialized agents enabled rapid prototyping with quality gates