B2B SaaS | AI CRO Platform

SplitSense is an AI-powered CRO & A/B testing platform that automatically generates copy variations, runs experiments, and optimises your websites around the clock, no code, no guesswork.

UX redesign & visual direction for an AI-powered conversion optimization platform, bringing agent insights, analytics, and session recordings into a clearer, more consistent product experience.

Splitsense helps teams identify friction in website journeys and find opportunities to improve conversion. Its AI agent, analytics, and session recordings each provide a different part of the picture. The design challenge was to help users move between those perspectives without losing context.

I worked on a UX redesign of the Agent Dashboard, Analytics, Navigation and Session Recordings experiences. Alongside the core flows, I defined a new visual language and a reusable design system to make the product feel like one connected platform.

The agent can surface findings continuously; the dashboard needs to help people understand which ones matter. I redesigned the view to make agent activity and emerging opportunities easier to scan, with a clearer distinction between information to monitor and items that call for a decision.

The aim was to make AI output useful in the flow of work: a starting point for investigation, with a clear route into the supporting evidence.

Conversion data becomes harder to use when every metric competes for attention. I reworked the analytics experience around a stronger information hierarchy, giving key performance signals prominence while keeping the detail available for deeper analysis.

The visual approach helps users read changes and trends before moving into individual journeys. It also creates a consistent way to present dense data across the platform.

Numbers can show where a journey breaks down; recordings help explain what happened. I redesigned the Session Recordings experience to make reviewing user behavior a clearer part of the investigation workflow.

This view gives teams a way to examine the behavior behind an analytics signal or agent finding, rather than relying on a summary alone.

The three experiences needed to work as one product. I established a modern visual direction and built reusable components for its data-heavy interface, creating consistency across layouts, information hierarchy, and interactive elements. The system built with variables and color tokens to centralize visual decisions, support scalable theming, and make the system easier to maintain across products and states.

The system gave the team a foundation for extending the platform while keeping new screens aligned with the redesigned experience.

Over ten weeks, the project delivered redesigned experiences for the Agent Dashboard, Analytics, and Session Recordings, supported by a reusable visual system.

The founder’s feedback highlighted that the UI met the team’s brief and addressed design problems they had been facing for some time. The result is a clearer foundation for a product that asks users to move between AI recommendations, quantitative data, and real user behavior.