Wright
Wright
Orchestrating AI agents For Smarter Data Exploration
Orchestrating AI agents For Smarter Data Exploration
Time
Time
Time
Jul 2025 – Sep 2025
Jul 2025 – Sep 2025
Jul 2025 – Sep 2025
Role
Role
Role
Product Designer
Product Designer
Product Designer
Team Size
Team Size
Team Size
1 designer + 6 engineers
1 designer + 6 engineers
1 designer + 6 engineers
OVERVIEW
OVERVIEW
Introduction
Introduction
Introduction
Wright is an AI-powered workspace that allows users to instruct the system to perform analytical tasks, compare documents, and generate reports. Wright automatically builds and runs specialized agents to complete each task, enabling large-scale document analysis across multiple files and formats.
Wright is an AI-powered workspace that allows users to instruct the system to perform analytical tasks, compare documents, and generate reports. Wright automatically builds and runs specialized agents to complete each task, enabling large-scale document analysis across multiple files and formats.
Wright is an AI-powered workspace that allows users to instruct the system to perform analytical tasks, compare documents, and generate reports. Wright automatically builds and runs specialized agents to complete each task, enabling large-scale document analysis across multiple files and formats.
Responsibilities
Responsibilities
Responsibilities
Interface design, stakeholder communication, and MVP delivery.
Interface design, stakeholder communication, and MVP delivery.
Interface design, stakeholder communication, and MVP delivery.
The Objective
The Objective
The Objective
To design a unified AI workspace that could create and manage AI agents, analyze documents across multiple sources, and present results in a transparent, easy-to-understand interface.
To design a unified AI workspace that could create and manage AI agents, analyze documents across multiple sources, and present results in a transparent, easy-to-understand interface.
To design a unified AI workspace that could create and manage AI agents, analyze documents across multiple sources, and present results in a transparent, easy-to-understand interface.
The Outcome
The Outcome
The Outcome
Delivered the MVP that led to 1st place in a provincial-level government competition and validation for broader institutional deployment.
Delivered the MVP that led to 1st place in a provincial-level government competition and validation for broader institutional deployment.
Delivered the MVP that led to 1st place in a provincial-level government competition and validation for broader institutional deployment.
PROBLEM
PROBLEM
PROBLEM
How Might We Simplify Analysis Through Agentic AI?
How Might We Simplify Analysis Through Agentic AI?
How Might We Simplify Analysis Through Agentic AI?
Analysts and organizations often deal with massive amounts of data scattered across spreadsheets, reports, and APIs. Much of this work still relies on manual effort. This not only slows decision-making but also increases the likelihood of human error.
Analysts and organizations often deal with massive amounts of data scattered across spreadsheets, reports, and APIs. Much of this work still relies on manual effort. This not only slows decision-making but also increases the likelihood of human error.
The challenge was to design an environment where users could interact naturally with AI and manage multiple analysis processes,
The challenge was to design an environment where users could interact naturally with AI and manage multiple analysis processes,
GOAL
GOAL
GOAL
Integrate AI-Powered Smart Glasses Into Daily Police Operations
Integrate AI-Powered Smart Glasses Into Daily Police Operations
Integrate AI-Powered Smart Glasses Into Daily Police Operations
The goal was to design an AI workspace that transforms natural language requests into automated data analysis through specialized AI agents.
The goal was to design an AI workspace that transforms natural language requests into automated data analysis through specialized AI agents.
The goal was to design an AI workspace that transforms natural language requests into automated data analysis through specialized AI agents.
CASE-SPECIFIC DESIGN
CASE-SPECIFIC DESIGN
CASE-SPECIFIC DESIGN
Designing for Real-World Government Use
Designing for Real-World Government Use
Designing for Real-World Government Use
Wright was selected for a provincial-level government innovation competition, where it was applied to a real-world investigative scenario. The system was tasked with analyzing datasets provided by law enforcement, identifying relationships between entities, and generating analytical reports through its AI agent framework.
Wright was selected for a provincial-level government innovation competition, where it was applied to a real-world investigative scenario. The system was tasked with analyzing datasets provided by law enforcement, identifying relationships between entities, and generating analytical reports through its AI agent framework.
Wright was selected for a provincial-level government innovation competition, where it was applied to a real-world investigative scenario. The system was tasked with analyzing datasets provided by law enforcement, identifying relationships between entities, and generating analytical reports through its AI agent framework.
Wright connects to the law enforcement API to automatically begin the data analysis process. It generates specialized AI agents to clean, filter, and identify suspicious individuals within the dataset.
Wright connects to the law enforcement API to automatically begin the data analysis process. It generates specialized AI agents to clean, filter, and identify suspicious individuals within the dataset.
Wright connects to the law enforcement API to automatically begin the data analysis process. It generates specialized AI agents to clean, filter, and identify suspicious individuals within the dataset.
Generating Reports
Generating Reports
Generating Reports
Users can then upload external files like reports or spreadsheets to expand the analysis. As the investigation progresses, Wright continues creating new agents to complete user-assigned tasks, ultimately producing relationship graphs and markdown reports summarizing individuals and groups.
Users can then upload external files like reports or spreadsheets to expand the analysis. As the investigation progresses, Wright continues creating new agents to complete user-assigned tasks, ultimately producing relationship graphs and markdown reports summarizing individuals and groups.
Users can then upload external files like reports or spreadsheets to expand the analysis. As the investigation progresses, Wright continues creating new agents to complete user-assigned tasks, ultimately producing relationship graphs and markdown reports summarizing individuals and groups.
DESIGN HIGHLIGHTS
DESIGN HIGHLIGHTS
DESIGN HIGHLIGHTS
Integrating Agentic AI Into a Seamless Workspace
Integrating Agentic AI Into a Seamless Workspace
Integrating Agentic AI Into a Seamless Workspace
LESSONS LEARNED
LESSONS LEARNED
LESSONS LEARNED
Designing For Scalable AI Collaboration
Designing For Scalable AI Collaboration
Designing For Scalable AI Collaboration
Through Wright, I learned how to design systems that make complex, multi-agent processes feel simple and actionable.
Through Wright, I learned how to design systems that make complex, multi-agent processes feel simple and actionable.
Through Wright, I learned how to design systems that make complex, multi-agent processes feel simple and actionable.
This project deepened my understanding of how to structure workflows around AI collaboration, balancing automation with user control to make powerful technology feel intuitive and efficient.
This project deepened my understanding of how to structure workflows around AI collaboration, balancing automation with user control to make powerful technology feel intuitive and efficient.
This project deepened my understanding of how to structure workflows around AI collaboration, balancing automation with user control to make powerful technology feel intuitive and efficient.
Skyler's Portfolio
Skyler's Portfolio
Skyler's Portfolio