Token Usage Dashboard
Designing trusted visibility for a shift to usage-based pricing (pre-launch)
The company was transitioning from bespoke, manually estimated pricing to a usage-based model.
For this to work, customers needed to understand how their applications consumed tokens without exposing unreliable or misleading data.
The core design challenge was deciding what we could show without creating false confidence or mistrust.
The problem
I partnered with product, engineering, DevOps, and data to help define a shippable MVP.
My focus was on:
Clarifying user needs across business and technical stakeholders
Surfacing trade-offs early
Constraining scope to reliable signals that could ship
What I worked on
Narrowed the MVP to business users focused on planning and budgeting
Avoided granular breakdowns that looked precise but weren’t reliable
Designed around trends, remaining capacity, and planning signals not diagnostics
Treated alerts and deeper diagnostics as fast follows once infrastructure matured
Key Decisions
The result was a token usage dashboard that:
Centralized usage visibility in the admin experience
Helped business users answer “Am I on track?” and “Where should I pay attention?”
Reduced risk in adopting a new pricing model
Established a scalable foundation for future iterations
Outcome
Usage based pricing only works if customers trust what they see. This project helped bridge the gap between incomplete data and confident decision-making without over compromising accuracy.