Creating and implementing a web portal for managing high-quality synthetic training data through user research for a drone tech startup

ROLE
UX Designer

TIMELINE
January 2023 - May 2023

TEAM
4 Designers, 1 Project Manager

TOOLS
Figma
Figjam

SKILLS
User Research
Web Design
Interaction Design
Product Strategy
UX/UI Design

Project Overview
Forming.AI is a cutting-edge startup that provides synthetic imagery services for enterprise clients, such as Lockheed Martin and BNSF Railway. By analyzing image data and predicting potential asset damage, Forming.AI helps these organizations minimize downtime, reduce costs, and streamline operations. In pursuit of a more effective client experience, the team set out to create a web portal where clients can seamlessly share their visual assets, monitor project milestones, and receive valuable analytical outputs—ultimately enabling better decision-making and proactive maintenance strategies.

Problem Space
While Forming.AI’s underlying technology is highly sophisticated, its existing client communication and collaboration workflows posed several challenges. Currently, conversations and updates are spread across multiple channels (e.g., Slack, email), making it difficult to track feedback and actions in one place. Clients and technical artists lack a centralized system to review progress, milestones, or deadlines, leading to misalignment in expectations. Clients want an intuitive way to annotate, comment, and provide input on generated images and models, but current tools do not support real-time, collaborative feedback. Technical artists struggle to keep clients on schedule for feedback or approvals, as there is no built-in notification or task-management system.

Without a transparent portal, both clients and technical artists face recurring friction points, limiting the smooth delivery of Forming.AI’s services, which leads to the problem:

How might we design a high fidelity prototype web portal that allows our customers to share assets with us, manage their synthetic imagery projects, and find value from the services we are providing to them?

My role as Synthesis Lead on this project involved guiding the design team through the research, ideation, and synthesis phases. I coordinated user interviews, consolidated cross-functional insights, and oversaw the overall design direction to ensure alignment with both client and technical artist needs.

Final Product
The high-fidelity prototype culminates in a polished web portal where clients and Forming.AI technical artists can collaboratively manage synthetic imagery projects in a centralized environment.

Main Dashboard
The Main Dashboard allows users to view all active, completed, and inactive  projects. Users will be able to search and filter projects as well as have easy view to which projects have new updates.

Project Page
Each individual project page contains the project description, reference images, project timeline, and all relevant files It allows users to access reference images, mark them up, and leaving necessary feedback/comments.  

Progress Report
An individualized progress report page for each project in a graphical format. Contains a detailed timeline of progress with expected deadlines, bar graph of number of comments and images uploaded

Profile
Profiles will be divided into Admin and Restricted accounts. Admin accounts have the ability to assign and deassign members to different projects, and restricted accounts can only have access to their assigned projects.

The Process

Research

To gain a deeper understanding of the diverse needs and pain points in the current workflow, we conducted 30-minute interviews with representatives from Lockheed Martin, BNSF Railway, and Forming.AI’s own technical artist team.

Goals
Understand client expectations and desired outcomes from synthetic imagery projects.Identify existing pain points in communication and project management
Key Findings
Seamless, Organized System: Clients need a single hub for documentation, real-time annotations, and interactive communication.
Clear Project Timelines & Expectations: Better visibility into project milestones, deliverables, and deadlines fosters trust and efficiency.
Streamlined Conversations: Replacing scattered Slack and email threads with a structured portal would mitigate confusion and loss of information.

Goals
Pinpoint challenges that arise when translating client requests into actionable tasks.Explore how the portal can support artists in managing timelines, feedback, and deliverables.
Key Findings
Notification System: Artists need automated reminders prompting clients to provide feedback or approvals on schedule, ensuring project timelines stay on track.Centralized Documentation & Annotation: A unified feedback and annotation system reduces back-and-forth and clarifies design revisions.Expectation Management: Clarity around model mockups, image specifications, and project scope is essential to delivering high-quality work efficiently.

Synthesizing the research
After conducting user interviews with both Forming.AI’s clients and technical artists, our team organized the data via affinity mapping and journey mapping. This process allowed us to isolate recurring themes, uncover critical friction points, and develop clear insight statements to guide our design efforts.


Affinity Mapping

Using an affinity diagram, we clustered the raw findings into eight overarching categories—Organization, Communication, Documentation, Timeline, Administration of Accounts, Billing, AI Models, and General Roadblocks. Each category corresponded to a set of pain points and goals


Journey Map

To visualize how these pain points manifest in real time, we constructed two journey maps:
Client Workflow: Traced how clients initiate and track projects, communicate with Forming.AI, and provide or request image markups.
Technical Artist Workflow: Highlighted the internal process for receiving and interpreting client feedback, iterating on AI models, and delivering final outputs.
By mapping these end-to-end experiences, we pinpointed the most critical friction areas, such as uncoordinated communication channels and poorly defined deadlines, which often stall progress and diminish transparency.


Insight Statements

From our affinity and journey mapping, four core insights emerged that directly informed our subsequent design decisions:

1. Organization
There is a general lack of organization for files, and no standardized naming process. Locating and editing files is done manually, and there is no formal method of sending/receiving images for marking up.

2. Communication
Communication is done through email or slack, but there is no formal notification system for when changes are made to the project.

3. Documentation
Projects are discussed on zoom calls and there are few notes/documentation of these conversationsNeeds to be history system that shows previous annotations/mark ups for references

4. Timeline
Forming.AI and clients often have different timeline expectations, which can prolong the project and create administration challenges further down the line.There is also no notification system for deadline expectations.

These insight statements became the cornerstone for shaping a user-centric, efficient web portal. They informed subsequent ideation, where we focused on building robust file organization, centralized communication tools, thorough documentation features, and transparent timelines to meet the needs of both clients and technical artists.

Ideation
Our ideation process kicked off with a series of brainstorming sessions aimed at addressing the core pain points revealed through user research. We focused on streamlining project organization, enhancing communication, and providing transparent timelines.

Notification System: Alerts for project updates, model downloads, and feedback requests.
Tiered Account Types: Administrative vs. Restricted access, allowing project managers to assign or revoke access based on specific roles.
Visualized Timeline: A milestone-driven display of project progress, ensuring both Forming.AI employees and clients stay aligned on deliverables and deadlines.

Low-Fidelity Prototypes
Our first round of tangible designs took shape in simple wireframes and paper sketches. The goal was to validate layout decisions and user flows before committing to more detailed visuals.


Popup Frames: Guiding users through onboarding and critical steps within the interface.
Streamlined Profiles: Keeping profiles minimal with key information like contact details, subscriptions, and assigned projects.

Flexible Exporting: Ensuring users can easily retrieve images, masks, and bounding box data in various formats.
AI Module Placeholder: Reserving space for future integration of AI-driven features.

Main Dashboard: Segmented into three sections—Active, Inactive, and Completed—so users can filter and track projects more efficiently.
Project-Specific Pages: Containing requirements, markups, chronological feedback, and deliverables in a vertical timeline for quick reference.

Mid-Fidelity Prototypes
Building on the feedback from our initial sketches, we transitioned to mid-fidelity wireframes with more refined layouts and interactive elements.

Reflections
Taking on the role of Synthesis Lead for this project offered a transformative experience in bridging user needs with design solutions. Coalescing findings from both client and technical artist interviews underscored the importance of active listening, focused alignment, and transparent communication—all critical to ensuring that the final product reflects real-world challenges. By orchestrating user interviews, guiding the affinity mapping, and translating insights into actionable design decisions, I gained a deeper appreciation for systematic research and evidence-driven ideation.

This project also marked my first foray into UX design and prototyping with Figma. Learning the intricacies of building and iterating on low- to high-fidelity prototypes was both challenging and rewarding, reinforcing the value of collaborative feedback loops for refining user flows. Ultimately, seeing our research-driven insights take shape in the high-fidelity portal not only validated our approach, but also illuminated the potential of user-centered design to meaningfully enhance workflows. The experience has solidified my confidence in tackling future UX initiatives and highlighted the critical role that structured, empathetic synthesis plays in successful product development.