↚ Back Home

Province AI Dashboard: A UX Design Case Study

How we designed trust into AI-powered construction technology through user-centered design

Project Overview

Province needed to design an AI-powered dashboard for construction pre-construction workflows.

The challenge: construction professionals are deeply skeptical of AI, yet desperately need efficiency gains in their chaotic, multi-stakeholder environment.

The Design Challenge: How do we make AI reasoning visible and trustworthy while solving the information chaos problem in construction workflows?

My Role: Lead UX Designer

Timeline: 1 week

Deliverables:

• Interactive Prototype
• Complete Screen Specifications
• Full App Architecture

This project showcases end-to-end UX capability: from market research and strategic concept development to working prototypes and detailed implementation specifications.

Full breakdown of the research in the link below.

Interactive Prototype Flow 1: Project Setup & Scope Analysis

1. Initial Research

The Hidden Reality of Construction

Research revealed a construction industry in crisis:

• Construction professionals spend 35% of their time hunting for information
• Projects involve 12+ stakeholders generating 2,000-10,000 documents over months
• 52% of project rework stems from communication failures, costing $31 billion annually
• Only 50% of pre-construction deadlines are met

The AI Adoption Paradox

• Only 52% of construction professionals have used AI tools (vs 91% in financial services)
• 25.7% cite data privacy as the biggest AI barrier
• Yet each technology adoption increases revenue by 1.14%

💡 Key Insight: Construction professionals don't want "black box" AI. They need to understand why the system made specific recommendations—they want explainable AI that shows its work.

What They Actually Want vs. Tech Assumptions

What Construction Professionals Need:

• Data integration across their 11 different systems (28.6% top priority)
• Time savings through automation (23.2% measure success)
• Early risk detection (21.6%)
• Workflow integration, not replacement systems

What Tech Companies Build:

• Sophisticated AI capabilities
• Feature-rich platforms requiring significant training
• Replacement systems rather than augmentation tools

Research Conclusion: They don't want AI features—they want problems solved with transparent reasoning.

2. Ideation

The Reframe That Changed Everything

Initial Challenge: "How do we make AI recommendations user-friendly?"

Reframed Challenge: "How do we make AI reasoning visible and trustworthy while solving information chaos?"

Key Ideation Sessions

Information Architecture Solutions

How Might We eliminate the 35% of time spent hunting for information?

Generated Solutions:

• Smart Document Tagging: AI tags by trade, phase, and stakeholder relevance
• Contextual Information Delivery: Role-based dashboards
• Progressive Disclosure: Summary-first with drill-down capabilities

💡 Breakthrough: Organize by decision points and workflow stages, not document types—matching how construction professionals actually think.

AI Transparency Solutions

How Might We make AI recommendations trustworthy?

Generated Solutions:

• "Show Your Work" Interface: Step-by-step reasoning display
• Confidence Scoring: Rate AI certainty, highlight when human review needed
• Alternative Options: Always show 2nd and 3rd best choices with reasoning

💡 Critical Insight: Construction professionals are used to peer review. AI should explain like an experienced project manager would to a colleague.

Design Principles Established

1. "Augment, Don't Replace": Enhance existing workflows
2. "Show, Don't Just Tell": Make AI reasoning visible through interface design
3. "Field-First, Office-Compatible": Optimize for harsh jobsite conditions
4. "Trust Through Transparency": Every recommendation includes clear reasoning

3. User Personas

I developed three interconnected personas representing the pre-construction ecosystem:

Primary: "The Field General" - Project Manager

Mike Rodriguez, 42

• Manages 3-5 concurrent projects worth $2-15M each
• Skeptical of new software due to past disappointments
• Key Pain Point: Spends 35% of time hunting for information across multiple systems

"I need to see exactly why the AI recommended this subcontractor, not just trust that it knows best."

Secondary: "The Bid Whisperer" - Estimator

Sarah Chen, 34

• Excel power user responsible for scope breakdown
• Works under intense bid deadline pressure
• Key Pain Point: Manual scope extraction from 200+ page specifications

"I'm open to AI help, but it needs to integrate with my workflow and let me verify its reasoning."

Tertiary: "The Tech-Forward Owner" - Development Rep

David Park, 38

• Expects modern software experiences
• Needs informed decisions without deep construction expertise
• Key Pain Point: Gets oversimplified updates that don't help decision-making

"Show me why this subcontractor was selected and what alternatives were considered."

Critical Handoff Points

Rather than separate experiences, I designed for persona intersections:

1. Estimator → Project Manager: Scope packages and recommendations
2. Project Manager → Owner: Status updates and decision requests
3. Owner → Project Manager: Approvals and change requests

4. Information Architecture

The Strategic Shift

Traditional Approach: Organize information by document type

Our Innovation: Organize by decision points and workflow stages

Primary Navigation Structure

├── Dashboard (Home)
├── Projects
├── Contractors
├── Notifications
└── Profile/Settings

Role-Based Information Delivery

Same Data, Different Views:

• Project Managers: Mobile-first, quick decision tools, stakeholder communication
• Estimators: Detailed analysis, advanced search, export capabilities
• Owners: Executive dashboards, decision audit trails, performance tracking

Trust-Building Architecture

Every screen includes:

• AI Confidence Indicators: Visual trust signals
• Reasoning Display: Expandable "Show Your Work" sections
• Alternative Options: 2nd and 3rd choices always visible
• Source Citations: Direct links to originating documents

5. Core User Flows

I designed four critical flows where AI provides maximum value while building trust:

Flow 1: Project Setup & Scope Analysis

Goal: Get AI assistance breaking down project scope

User: Estimator (primary), Project Manager (secondary)

Key Innovation: Real-time processing transparency

• Document upload with clear AI expectations
• Live progress: "AI analyzing drawings pages 12-18, found 47 scope items"
• Side-by-side review: AI suggestions vs. source documents
• One-click export to existing estimating workflows

Flow 2: Contractor Discovery & Matching

Goal: Find qualified subcontractors with explainable logic

User: Project Manager (primary), Owner (approval)

Key Innovation: Reasoning-first recommendations

• "Based on similar projects, this contractor excels at complex MEP coordination"
• Historical performance data with project-specific context
• Alternative contractor analysis with trade-off explanations

Flow 3: Decision Review & Approval

Goal: Multi-stakeholder decision-making with audit trails

Users: All personas in approval chain

Key Innovation: Complete decision reasoning

• Full rationale documentation
• Alternative analysis and trade-offs
• Approval workflow with delegation options

Flow 4: Mobile Field Verification

Goal: Validate AI recommendations in real-world conditions

Context: Harsh jobsite, gloves, poor connectivity

Key Innovation: Offline-first field experience

• Glove-friendly interface (60px touch targets)
• High contrast for sunlight readability
• Intelligent sync when connectivity returns

User Flow Mapping

Total Experience: 4 core flows, 23 unique screens

• Project Setup Flow: 8 screens
• Contractor Discovery: 6 screens
• Decision Review: 5 screens
• Field Verification: 4 screens

6. User Screens

Mobile-First Design Philosophy

Primary Context: Outdoor jobsites with bright sunlight, gloved hands, poor connectivity

Design Specifications

• Device Target: iPhone 14/15 Pro (390x844px) and Android equivalent
• Touch Targets: 60px minimum for gloved use
• Typography: 16px minimum, 24px+ for critical information
• Contrast: 7:1 ratio for outdoor readability
• Connectivity: Offline-first architecture

Key Screen Examples

1. Dashboard Home

Purpose: Project overview and quick access to critical functions

Header (80px)
├── Province Logo
├── Notification Bell (with badge)
└── Profile Avatar

Status Cards (240px)
├── "3 Active Projects"
├── "2 Decisions Need Review" (red alert)
└── "AI Processing: 67% Complete"

Quick Actions (120px)
├── "Start New Project" (Primary CTA)
└── "Find Contractors"

Trust Elements:

• AI processing status with percentage
• "Last updated 3 minutes ago" timestamps
• Clear confidence indicators on recommendations

2. AI Processing Status

Purpose: Show real-time analysis progress, build trust through transparency

Progress Section (400px)
├── Progress Bar: "Analyzing Documents... 34%"
├── Current Task: "Extracting scope from drawings"
├── Time Estimate: "~3 minutes remaining"
└── Step-by-Step Status:
    ├── "✓ Document upload complete"
    ├── "🔄 Reading architectural drawings"
    ├── "⏳ Identifying scope items"
    └── "⏳ Matching trade requirements"

Innovation: No "black box" processing—users see exactly what AI is doing

3. Scope Review Dashboard

Purpose: Review and validate AI-identified scope items

Summary Cards (120px)
├── "47 Scope Items Found"
├── "87% Confidence Score"
└── "3 Items Need Review"

Scope Items List
├── Item with confidence indicator
├── Source citation: "Found in Spec 03000, Page 12"
├── Edit/Approve/Reject actions
└── "Show AI Reasoning" expandable

Trust Building:

• Confidence scores for each item
• Direct source citations
• Easy correction mechanisms
• Transparent AI reasoning

4. Contractor Profile Detail

Purpose: Detailed contractor evaluation with AI reasoning

Contractor Overview
├── Rating: "94% Match for this project"
├── "Why recommended" expandable section
└── Alternative contractors shown

Key Qualifications
├── Relevant project history
├── Licensing and insurance status
└── Performance on similar projects

AI Reasoning Display
├── "Excels at complex MEP coordination"
├── "Successfully completed 12 similar projects"
└── "Average 5% under budget, 2 days early"

Design System Components

Trust-Building Elements

• AI Confidence Dots: Green/Yellow/Red with percentages
• "Show Your Work" Sections: Expandable reasoning displays
• Source Citations: Clickable links to original documents
• Alternative Options: Always visible second choices

Field-Ready Accessibility

• Large Touch Targets: 60px for glove use
• High Contrast Colors: 7:1 ratio minimum
• Offline Indicators: Clear connectivity status
• Battery Optimization: 8+ hour field operation

Screen Flow Connections

All 23 screens connect through intuitive navigation:

• Bottom tab navigation for primary functions
• Contextual back buttons maintaining user mental model
• Quick actions accessible from any screen
• Emergency contacts always available

8. Results & Impact

Immediate Outcomes

• Clear MVP Definition: 23 screens with validated user needs
• Technical Roadmap: Mobile-first, offline-capable architecture requirements
• Trust Framework: Consistent AI explanation patterns
• Development-Ready: Complete prototype specifications
• Interactive Prototype: User flow 1 and 2 (will build a couple more)

Key Design Insights

1. Trust Through Transparency: Construction professionals will adopt AI if they can understand and verify its reasoning
2. Mobile-First Essential: Field usability wasn't optional—decisions happen on jobsites
3. Integration Over Innovation: Success came from enhancing existing workflows, not replacing them

Validation Strategy

Testing Scenarios:

• New project setup and scope analysis (10-minute completion goal)
• Contractor selection with reasoning understanding
• Field verification with offline functionality

Success Metrics:

• 90%+ task completion rate for primary flows
• 4.0/5.0+ user confidence in AI recommendations
• 85%+ field usability success rate

Next Steps

1. Full Workflow Prototype: Complete interactive demo of all user journeys using placeholder AI responses
2. Design Pattern Validation: Test UX approaches with construction professionals to validate workflow logic and interface effectiveness
3. Development Roadmap: Document technical requirements for AI implementation and system integrations for future partnerships
4. Market Feasibility: Validate business concept and user interest through stakeholder presentations and feedback

Real Project Implementation

If This Were a Live Project with Development Team:

1. Primary User Research: Stakeholder interviews across all 15+ user types, workflow shadowing during actual preconstruction processes
2. Prototype Validation: Usability testing with construction professionals using working demos in field conditions
3. Technical Integration: API assessment with Procore/Autodesk teams, offline architecture requirements validation
4. Pilot Program: 2-week trials with 3-5 construction firms, iterative refinement based on real usage data

Project Impact

The design establishes a new standard for construction AI: transparency-first, field-ready, and workflow-integrated.

The key breakthrough: Rather than making AI more user-friendly, we made AI reasoning visible and trustworthy, exactly what construction professionals needed to embrace this transformative technology.

Nuvra Health →