Case study · Moment design challenge
One investigation workflow for ops clarity
Problem in one sentence: Small and midsize companies cannot easily understand spending and performance across cloud, subscriptions, AI APIs, and automated workflows — so anomalies become end-of-month surprises.
Users & context
Founder/CFO, Ops manager, and Technical lead at a fictional company (Northline Commerce). Sample data only — portfolio concept, not a billed SaaS.
Role
Self-initiated product design + front-end build by Kiril Mironyuk: brand system, investigation IA, and clickable Next.js prototype.
Research & limitations
Competitive teardown of FinOps / SaaS spend tools and informal interviews with ops practitioners. No production customer dataset; all figures are labeled sample data.
Opportunity
Narrow to one excellent path — alert → spike → cause → recommendation → approval → tracking — instead of a decorative dashboard or ad campaign.
Success criteria
A reviewer can complete the investigation in under two minutes, understand the business value in 30 seconds, and see empty/loading/success states, audit trail, and developer notes on the prototype.
Primary journey
- Step 1
Executive summary
Surface the important change — an AI spend spike — without a wall of charts.
- Step 2
Cost spike
Show when and where the anomaly began with a dense daily series and table.
- Step 3
Probable cause
Connect evidence to the Support summarizer and GPT-4o default routing.
- Step 4
Recommendation
Explain action, risk, confidence, and estimated monthly impact.
- Step 5
Approval
Assign ownership and preserve an auditable decision trail.
- Step 6
Tracking
Show whether the change produced expected savings after approval.
Visual system
Navy base, cyan signal, green savings. Glass surfaces for density without clutter. Pulsing signal nodes carry continuity across marketing and product without repeating one layout.
Demo login: demo@mirotech.io / ops-demo · Sample data labeled throughout
Reflection & next steps
Depth on one workflow beats a wide marketing kit. Next: live connectors, shared audit storage, and stronger chart accessibility (data tables already accompany the spike sparkline).