Due to confidentiality, this image is merely representative. Further details can be shared upon case walkthrough scheduled via hello@matildedosreis.com
01 / Context
During a recurrent meeting circles in which user representatives seat, one day I carefully listened to a stakeholder describing a manual workflow that is crucial to their projects audits - "it is a time consuming and painful old process" - he said. Right after that meeting, I set up a sync with product owners to discuss this need and that was the very start of what became one of the major exploration topics included on the product development roadmap forecast. The automation of this legacy process was indeed considered as a part of the product vision.
02 / Problem statement
The current document creation process across sales and quality control teams is inefficient and error-prone, requiring users to switch between multiple tools and recreate documents when issues arise. This leads to poor data quality, increased operational effort, and reduced team productivity.
03 / Opportunity mapping
Translating research findings into actionable opportunities — each pain point was mapped to a concrete design direction and potential measurable impact.
| Area | Current Pain Point | Insight | Opportunity | Potential Impact | Example Metrics |
|---|---|---|---|---|---|
| Workflow Efficiency | Users spend ~5 hours per document | Process is fragmented and manual | Streamline into a unified document creation flow | Reduced task time, faster delivery | ↓ Task time (5h → <2h) |
| Tool Fragmentation | Users switch between 3 tools | Context switching increases cognitive load and errors | Create a single source of truth / integrated platform | Lower mental load, fewer mistakes | ↓ Tool switching, ↓ drop-offs |
| Error Rate | 1 in 5 documents needs correction | Lack of validation during creation leads to rework | Introduce real-time validation & error prevention | Improved data quality, less rework | ↓ Error rate (20% → <5%) |
| Rework Effort | Users recreate documents from scratch | No versioning or editing flexibility | Enable inline editing + version control | Reduced duplication of effort | ↓ Rework time, ↑ efficiency |
| Data Quality | Poor data consistency across teams | No standardized inputs or validation rules | Standardize with structured templates & rules | More reliable outputs, better reporting | ↑ Data accuracy |
| Collaboration | Sales & QA workflows are disconnected | Lack of shared visibility across teams | Build collaborative workflows (handoffs, status tracking) | Smoother cross-team coordination | ↓ Handoff delays |
| User Confidence | Users don't trust the process (errors, duplication) | System doesn't provide feedback or reassurance | Add progress indicators + validation feedback loops | Increased confidence and adoption | ↑ Task success rate |
| Operational Cost | High time investment + duplication | Inefficiencies scale with volume | Automate repetitive tasks (auto-fill, reuse data) | Cost savings, scalability | ↓ Cost per document |
04 / Business goals
Each goal was previously aligned and agreed between the product team and stakeholders before any design work began.
Single point of truth
Build a unified system to ensure data quality across all departments.
Reduce manual effort
Automate content generation to eliminate repetitive document creation.
Automate distribution
Replace manual document distribution with automated delivery flows.
Prevent manual errors
Remove error-prone steps that forced users to restart documents from scratch.
Standardise processes
Establish a consistent working process across all teams and departments.
Track version changes
Provide full visibility into document version history and change tracking.
04 / My role
Lead designer from discovery to handoff — research, prototyping, testing, and stakeholder alignment.
User research
7 contextual interviews · opportunity mapping · user journey mapping
Stakeholder alignment
Weekly reviews with CFO and business responsible
Prototyping & testing
3 iteration rounds · 6 participants each
Dev handoff
Interaction specs · embedded 2 sprints
04 / How I measured success
90-day pre-launch baseline. Measured at 30 / 60 / 90 days post-ship.
Task time completion
30 min
5 hrs → 30 mins
Backend timestamps per publishing
Correction rate
−28%
21% → 15% of documents
Error flags in approval log
Task success
94%
Up from 61% in round 1
Usability test, round 3
Approver NPS
+41
Baseline was 12
In-product pulse survey
05 / Trade-offs
All trade-off decisions were the result of cross-team collaboration and discussion, involving design, engineering, product, and business stakeholders throughout the process.
Single point of truth
Centralised all document data into one system, enforcing a single source of record across departments.
Teams lost local copies and personal spreadsheets they relied on. Required change management to migrate legacy data.
Automated content generation
Templates auto-populate fields from existing data sources, eliminating manual document creation from scratch.
Reduced flexibility for edge cases. Users with non-standard documents needed a separate manual override path.
Automated distribution
Documents are automatically routed to the right recipients on publish — no manual forwarding or email chains.
Removed ad-hoc distribution control. Teams had to reconfigure distribution rules upfront, adding onboarding time.
Error prevention
Inline validation and mandatory field checks block submission until data quality requirements are met.
Stricter validation created friction for power users who wanted to save drafts mid-way through incomplete forms.
Standardised process
A unified workflow enforced across all teams replaced department-specific variations, ensuring consistency.
Departments with legitimate process differences required exceptions, increasing configuration and governance complexity.
Version tracking
Full version history with change attribution gives teams an audit trail and the ability to compare revisions.
Storage and performance overhead increased. Surfacing version diffs clearly without overwhelming the UI required extra design iteration.