Ryan Schmidt
← Back to all case studies

AI‑Powered Custom Dashboards

Let every user see the metrics that matter via customizable widgets and AI reporting (beta testing).

  • +71% dashboard engagement (vs. baseline)
  • 62% created multiple dashboards
  • 3.0 median widget interaction depth
My role
Lead Designer & Acting PM. Responsible for problem framing, IA, prototyping, usability validation, launch narrative. Collaborated with engineers (5 months, 4 devs, remote).

Problem

  • Users: “The current dashboards aren’t useful. Give us more metric cards and more dashboards so we can actually manage our work.”
  • Stakeholders: Similar ask; Enterprise users are going to need more dashboards, for change orders, service, etc.

Goals

  • Enable personalized dashboard creation for every user
  • AI-generated executive reports in plain language
  • Demonstrate value instantly in demos, boost sales win-rate
  • Drive ongoing product engagement
  • Power future Insights+ product with usage metrics

Research Highlights → Design Implications

  • Methods: 10 interviews (6 users, 4 stakeholders), competitive analysis
  • Key behavioral findings: Users often disengaged from default dashboards → prioritized customization, more meaningful metrics
  • Internal insights: Sales needed "instant value" → implemented role‑based templates + easy customization
  • Key request: Personalization and AI-driven reporting were recurring user asks

Design Approach & Decisions

Dashboard builder with drag-and-drop widgets
Users can create additional dashboards and choose from a variety of customizable widgets, focused on all aspects of their business.
AI recommendations panel suggesting metrics with accept/edit controls
Drag-and-drop widgets with resize, show/hide information, change dataviz style, and snap-to-grid.
Role-based template gallery for quick start
Dashboards are highly-customizable and specific to the user, or their organization.
Analytics view showing real-time visualizations
AI generates a report PDF for deeper dives and a What is Going Right and What is Going Wrong section.

Prototyping & Validation

Interactive Figma prototype tested with target users on role-based templates, customization interactions, and generative reporting.

  • Role-based templates: 100% of participants agreed that these were an improvement over existing and easy to use
  • Customization: Users loved the ability to create multiple dashboards and add, remove, customize, or resize widgets.

Business Value & Lessons

  • What appears to be a minor UX improvement actually reshapes core workflows and outcomes. This is what enterprise users are looking for based on feedback given to our sales team, CS team, and directly to me by existing users. This is also a great entrypoint in enabling some industry-defining AI capabilities. As the suite expands, we eliminate need for outside BI tools. This saves our users money and keeps them logged into our software more often and for longer periods.

Roadmap/Next

  • While large users wanted some additional metrics to be tracked, we're planning on adding that prior to the MVP as well as in future iterations. In general, users loved this feature, and we're excited to launch it in our 2025 December release.
  • Add cross‑dashboard correlations, richer natural‑language queries (“compare QoQ”), and team‑shared templates. AI assistance in recommending additional widgets based on the user's role and interaction patterns.