Cognitive Procurement · Spend Analysis
Spend Analysis
Simplifying complex spend exploration through a more intuitive filtering experience.
Role
UX Designer
Product
Digitate Cognitive Procurement
Module
Spend Analysis
Focus
Filtering & categorization experience
Due to client confidentiality, certain product details and data have been simplified for this portfolio case study.
Project Snapshot
Product
Cognitive Procurement
Module
Spend Analysis
Focus
Filtering / categorization
Role
UX Designer
Scope
Filter interaction & UX
01 — Product Context
Cognitive Procurement
Cognitive Procurement is an AI-powered, SaaS-based spend intelligence platform by Digitate that helps enterprises make smarter procurement decisions. It connects with the Procure-to-Pay (P2P) ecosystem to provide visibility into organizational spend, identify opportunities, and continuously monitor transactions for risks and anomalies. The platform brings together capabilities across Spend Analysis, Spend Classification, Supplier Analysis, and Contract Analysis—helping procurement teams move from fragmented spend data to faster, more informed decisions. This case study focuses on the filtering experience within Spend Analysis—making it easier for users to navigate complex spend data and quickly find the information they need.
What Is Spend Analysis?
Spend Analysis is a core module within Digitate’s ignio Cognitive Procurement platform, designed to help organizations understand, control, and optimize their procurement spend. It brings together purchasing data, cleans and classifies it, and enables users to analyze spending across categories, suppliers, departments, and other dimensions. By making spend patterns, anomalies, pricing variations, and potential savings opportunities easier to identify, the module supports better sourcing decisions, improves spend visibility, and helps organizations reduce leakage and drive procurement efficiency.
02 — The Challenge
The team believed the filter needed a complete redesign. The real challenge was discovering whether that was true.
The existing filter offered simple dropdown-style selection. For complex, large-scale spend data, it didn't give users enough visibility or clarity — or confidence about what they'd actually selected. Product Managers and Product Owners were frustrated, and a full redesign of the filter flow was on the table.
Option A
Complete Redesign
Bigger structural change — but more dev effort, more dependencies, longer cycle, and possibly unnecessary if the real problem is narrower.
vs
Option B
Focused Improvement
Targets the actual interaction problem — lower complexity, preserves the existing system, solves what's really broken.
The real question wasn't "how do we redesign a filter?"
— it was "how might we make complex filtering clearer without redesigning more of the system than necessary?"
03 — The Existing Experience
Why this wasn't a simple filter
Large transaction dataset
→
Multiple categorization levels
→
Specific / niche selection
→
Confidence in what's selected
04 — Exploring the Broader Redesign
The initial assumption
At first, the logical direction seemed obvious: redesign the complete filter flow. I explored conceptually how a broader redesign might work — but that exploration exposed a more important question than any specific layout.
Redesign everything?
↓
Would it actually solve the user's problem?
↓
I needed evidence, not assumptions
↓
Talk to the actual user
Conceptual representation — no low-fidelity wireframes exist from this stage of exploration.
05 — The Research Pivot
Talking to the actual user
Rather than assuming the entire flow needed to change, I asked the Product Owners to connect me with someone who actually used the filter day to day. That conversation ran roughly one to two hours, and it became the turning point for the project.
Stakeholder discussions
+
Actual user conversation (~1–2 hrs)
+
Workflow investigation
=
Real understanding of the problem
What I learned
01
The entire workflow did not need to change.
02
The biggest issue was visibility and clarity during filtering, not the flow itself.
03
Users needed a better way to navigate specific/niche selections within huge result sets.
04
The solution needed to work within the existing system, not add backend complexity.
"Redesign the filter flow."
↓
"Improve the filtering interaction where users actually struggle."
The problem became smaller — and the solution became stronger.
06 — Design Principles
07 — The Final Solution
A filtering experience built for confidence, not just selection
The filter now walks the user through categorization level → search and bulk-select within huge result sets → materials → a full review of every selection → save for reuse. Each step keeps the user oriented in what could otherwise be a 50,000-row list.
Step 01 — Opening the Filter
A clear starting point
A dedicated "Your Selection" panel sits alongside every filter tab from the very first interaction.
Users always know where their selections will appear before they make one.
Step 02 — Choosing a Category Level
Selection confirmed before moving on
Category Level selections (e.g. "Category Level 4") are checked, then immediately reflected as a removable tag in "Your Selection."
Visible confirmation at every step — no wondering if a choice actually registered.
Step 03 — Searching & Selecting at Scale
Making 50,000 results navigable
A "Search Within Selected" field plus a "Select all searched results" bulk action, with a live count of the result set and how many are selected.
This is the core niche-selection problem solved — users no longer scroll a flat list to find and check items one by one.
Step 04 — Refining Within a Selection
Searching inside what's already selected
The same search-and-bulk-select pattern works recursively — inside an already-selected group, not just the full list.
Supports exactly the "find a specific/niche subset" scenario the original interview surfaced as the real pain point.
Step 05 — Layering Additional Dimensions
Selections accumulate, not replace
Materials filtering follows the same pattern, and "Your Selection" now tracks Category Level, Categories, and Materials in parallel.
One consistent mental model across every filter dimension — learn it once, use it everywhere.
Step 06 — Reviewing Every Selection
A single place to see everything before applying
A dedicated review modal shows every individual selection at once, searchable, instead of a collapsed "+492 more" tag.
This is the direct answer to the original problem — "confidence about what users were selecting" before running the analysis.
Step 07 — Saving & Reusing Filters
Complex filters, saved for next time
Any combination of selections can be named and saved, then reapplied or edited later from a "Saved Filters" list.
A complex, precisely-built filter — like the 56-category selection above — never has to be rebuilt from scratch.
08 — Key Design Decisions
What changed, and why
Decision
Persistent "Your Selection" panel
Reasoning: Selections were previously invisible until applied.
Impact: Users see exactly what they've built, at every step.
Decision
Search + bulk-select within results
Reasoning: Manually checking items in a 50,000-row list was the core friction point.
Impact: Niche selections go from minutes of scrolling to a single search.
Decision
Nested search within a selection
Reasoning: Users needed to refine a broad selection further, not just search the full list again.
Impact: Supports progressive narrowing without losing prior work.
Decision
Full-screen review modal
Reasoning: A collapsed "+492 more" tag isn't real visibility.
Impact: Users can verify every single selection before committing.
Decision
Save & reuse named filters
Reasoning: Complex filters were being rebuilt from scratch every session.
Impact: Recurring analyses become one click, not a rebuild.
Decision
Worked within the existing system
Reasoning: A full flow redesign would have meant significant backend rework for a narrower problem.
Impact: Meaningfully better UX without unnecessary system complexity.
09 — From Assumption to Direction
Old Assumption
Complete flow redesign
Final Direction
Focused filter experience improvement
No authentic old-filter screenshot was available for this case study, so this comparison represents the evolution of the design direction rather than a literal before/after UI.
10 — Outcome
What the focused direction achieved
✓
The need for a complete redesign was avoided.
✓
The solution focused on the actual user problem, not assumptions.
✓
The filtering experience became clearer and more visually understandable.
✓
The final direction aligned better with real user scenarios than the initial assumption.
11 — Reflection
What this project taught me
Don't assume the scope of a redesign before understanding the user's actual workflow.
A smaller, more focused solution can be the stronger product decision.
UX isn't only about improving the interface — it's about identifying what should and shouldn't change.
Good enterprise UX means balancing user needs, product goals, and implementation realities.
Talking to the actual user can invalidate assumptions made in stakeholder discussions — and it's always worth the hour.
See the interaction in action
Watch the Walkthrough
Explore the Prototype
Links are placeholders — to be connected once the prototype/recording is provided.
Next Case Study
