Process intelligence for Fortune 500s — but the data lived in siloed tools no designer had mapped before.
Soroco builds process intelligence software for Fortune 500 companies — tools that analyze how employees actually work versus how they're supposed to work. The data comes from enterprise systems: ERP, CRM, email, calendar. The insight is in the patterns.
When I joined in 2022, the platform had strong backend capability and a frontend that read like an analyst's spreadsheet. The primary users were operations managers and process improvement leads — people who understood the data intimately but spent hours every week just formatting it for stakeholder presentations.
My mandate was to redesign the core workflow mapping and analysis interface — a tool that Fortune 500 teams used to document, measure, and improve their business processes.
Process intelligence for Fortune 500s — but the data lived in siloed tools no designer had mapped before.
The process intelligence data was rich and multi-layered. The users were sophisticated — they didn't need to be protected from complexity. But the existing interface had no hierarchy: every data point was presented at the same visual weight. The result was an interface that looked complex even when the insight it was surfacing was simple.
The second problem: enterprise operations work is collaborative. Multiple stakeholders needed to view and annotate the same workflow map. The platform had no collaboration layer — process maps were exported to PowerPoint and shared over email, breaking the connection between the data and the insight.
The original interface: all data at the same visual weight. Finding the signal required knowing the data already.
The workflow data was stored in a proprietary graph format with no standard visualization library. Building an accurate process map visualization required close collaboration with engineering to expose the right data structures to the front end.
Soroco's customers are Fortune 500 companies with long procurement cycles. UI changes had to go through customer advisory board review, which added 4–6 weeks to any change cycle. I had to design for a validation process, not just a shipping process.
Redesign the workflow canvas for semantics, not just aesthetics
The original workflow canvas used generic diagramming conventions — boxes and arrows that could have come from any Visio template. We redesigned it around the actual semantics of process intelligence: activities, handoffs, wait states, and exceptions each had distinct visual treatments that made the process structure immediately readable without prior training.
Semantic visual grammar: four node types, each immediately distinguishable at a glance.
Separate the map from the analysis
We created a split-view: the process map on the left, the insight panel on the right. Clicking any activity or handoff in the map surfaced its performance data — time spent, frequency, deviation from standard — in the panel. This preserved the full map while making individual insights accessible without context-switching.
The insight panel: click any node, see its performance. No more cross-referencing with data exports.
The insight panel became the highest-rated feature in post-launch surveys. Users had been manually cross-referencing the map with exported data before this existed — sometimes spending 2+ hours on a single presentation.
We designed — and prototyped — an AI layer that would automatically identify "problem nodes" in the process map: activities with high deviation or long wait times, surfaced in priority order. We paused it. Enterprise ops teams have strong opinions about where problems are. An AI that confidently identifies a different set of problems can undermine the designer's credibility with their stakeholders before they've had a chance to build context.
Gartner ranking in process intelligence category
2023
Fortune 500 clients onboarded during redesign period
2022–2023
reduction in process mapping time
post-redesign baseline
The redesigned platform contributed to Soroco achieving a #1 Gartner ranking in the process intelligence category in 2023. Six Fortune 500 clients onboarded during the period the redesign was active. Process mapping time — time from data ingestion to a shareable workflow map — dropped by 40%.
The feedback that surprised me most: "Onboarding sessions got shorter." That wasn't a metric we tracked. It meant the interface was becoming self-explanatory.
Working in enterprise software taught me that adoption is a design problem, not a training problem. If the interface requires a training session, the design has failed. The most important thing I built at Soroco wasn't the visual redesign — it was the insight panel that made the data self-explanatory.
I also learned that designing for sophisticated users is harder, not easier. They have high standards, strong mental models, and the vocabulary to tell you exactly why something is wrong. That's a gift — use it.
