Insphere AI Platform
UX for an AI platform that designers and engineers can actually use
The Challenge
Most AI platforms fail because they're designed by engineers for engineers. Insphere needed a UX that made powerful AI capabilities accessible to designers, product managers, and non-technical users — without losing depth for power users.
The Process
Led a 3-week discovery phase: user interviews with 18 designers and PMs across the UK and US on their existing AI tool usage and pain points
Identified the core UX problem: AI platforms bury their most powerful features behind technical config screens that non-technical users never reach
Designed a progressive disclosure system: simple defaults with one-click AI execution for 80% of use cases, and full parameter control for power users
Built a comprehensive design system — 120+ components — that all future AI features are built on, ensuring visual and interaction consistency
Integrated AI tool flows into existing design workflows rather than forcing separate AI sessions — reducing context-switching by 50%
Architected the frontend with Next.js and TypeScript, improving page load times by 30% through component optimisation
The Outcome
The redesigned platform saw a 20% increase in user engagement. Iteration cycles for design teams using the platform dropped by 50%. The design system is now used across all Insphere product lines.
12+ years building digital products at the intersection of code and design. Senior Design Engineer at Insphere AI, UX Delivery Lead at ReloadUX, and Lead Designer at Tkxel. Based in Lahore, Pakistan.