Atoti UI Redesign

Enterprise UX • Activeviam

OVERVIEW

AtotiUI is a web-based enterprise analytics platform used by financial institutions for complex data exploration — traders, risk managers, and FP&A analysts building dashboards from large OLAP cubes.

When I joined ActiveViam in 2021, I inherited a product with strong analytical power and a fragmented interface. Over the first two years I led a systematic UI redesign focused on navigation architecture, workspace layout, and a no-code styling system — while establishing the design foundation the product still runs on.

MY ROLE

Sole UX/UI designer

COLLABORATION

Product managers, front-end engineering team

TIMELINE

2 years (ongoing)

THE PROBLEM

The original interface was built around drawer-based navigation — tools, filters, fields, and widgets each lived in separate panels that opened and closed over the canvas. The result was constant context switching: a user building a dashboard had to open the fields drawer, close it, open the widget drawer, close it, open filters. The canvas was never fully visible and the workspace never felt stable.

Two types of users made this harder to solve:

USERS

BUILDERS

Analysts, quants, and technical researchers who understand cube structures and need access to measures, hierarchies, and MDX tools. They were opening and closing drawers constantly.

EXPLORERS

Traders, risk managers, and financial analysts focused on answering business questions. They struggled to understand which drawer held what, and why customization required writing MDX queries at all.

The interface was optimized for neither.

How might we give both user types persistent access to the tools they need, without cluttering the analytical workspace?

WHAT CHANGED

1. FROM DRAWERS TO PANELS

The biggest structural change was replacing the drawer system with a persistent panel layout.

Before

After


Rosi Solano PortfolioRosi Solano Portfolio

LAYOUT AND IA IMPROVEMENTS

  1. NO-CODE STYLING SYSTEM

Conditional formatting, date formatting, and table styles previously required writing MDX queries, a significant barrier for non-technical explorers.

I introduced a visual styling editor in the TOOLS panel that makes these options directly accessible. Advanced users can still write MDX, but the interface no longer requires it.

Conditional formatting, date formatting, and table styles previously required writing MDX queries, a significant barrier for non-technical explorers.

Before: Conditional formatting required editing raw JSON:

After: Same rule, now set visually:

OUTCOME

The redesigned interface shipped across versions 5.1 to 5.3, adopted by 33 financial clients including BBVA, Commerzbank, HSBC, and Santander. The panel-driven layout became the product's architectural foundation, every feature since, including the AI analytics work, is built on top of it.

The design system reduced inconsistency across the product and gave engineering a shared language for implementation. What started as a navigation fix became the scaffolding for everything that followed.