Calix MVP

I designed Calix from zero to product-market fit—an AI-powered demand planning platform that replaced weeks of Excel work with same-day forecasts. What started as an MVP became iSystems' flagship product, growing from 3 pilot customers to 20+ enterprises with R$5M+ ARR.

Demand planners at manufacturing companies spent 1-2 weeks per cycle generating forecasts in Excel. The process was error-prone, impossible to scale, and left no time for strategic analysis.

• 100% of interviewed users relied on spreadsheets
• 60% used additional tools on top
• Trust in AI was low—users needed to see reasoning, not just output

Lead Product Designer (solo) — End-to-end ownership

I was the only designer on a team with PM, Tech Lead, and AI Specialist. I owned the entire design process: research with 30+ industrial planners, journey mapping, hypothesis validation, information architecture, interaction design, UI design, prototyping, and design system foundations.

• Conducted 30+ interviews with demand planners across industries
• Mapped complex planning workflows (4+ hour tasks per forecast cycle)
• Identified that AI trust depends on transparency—users need to understand why the model made predictions
• Validated that time savings (not features) was the primary value driver

The core insight was that planners don't just need a number—they need confidence in that number.

I designed the interface around three principles:

• Radical transparency: Show model inputs, assumptions, and confidence intervals
• Familiar mental models: Mirror Excel workflows to reduce learning curve
• Progressive disclosure: Simple overview first, drill-down for power users

• Dashboard-first: Show forecast vs. actual at a glance, not buried in menus
• Editable predictions: Let planners adjust AI output and see downstream impact
• Audit trail: Every change logged, every assumption visible
• QuickSight integration: Chose simplicity over custom dashboards for MVP speed

Selling time beats selling features. When we pitched "forecast in 1 day instead of 1 week," customers immediately understood the value. Early on, we focused too much on model accuracy metrics that planners didn't trust. Shifting focus to transparency and time savings unlocked adoption.