UX Designer Who Builds Trust.
5 years designing complex systems — healthcare AI, enterprise platforms, and design systems — where clarity, trust, and human judgment are non-negotiable.
Shaik Shekath Ali
I'm a UX / Product Designer with 5 years of experience working on complex, high-stakes products where design decisions carry real consequences — clinical workflows, enterprise systems, and AI interfaces.
My approach starts with understanding the system before touching the interface. I map stakeholder needs, audit existing workflows, and find the moments where friction creates the most damage — then design the minimum intervention that creates the maximum clarity.
I'm comfortable navigating ambiguity, constraints, and competing priorities. Whether working with clinicians, engineers, or executives, I translate complexity into decisions that stick.
Selected Work
HIP One: Designing for Trust in Healthcare AI
Transforming fragmented medical reviews into a unified, HIPAA-compliant decision workflow — where AI accelerates the work, but humans remain accountable for every outcome.
Stakeholder Interviews
Aligning regulatory non-negotiables and compliance requirements before any design work begins.
Contextual Inquiry
Shadowing manual PDF and EHR workflows in real clinical environments to understand true pain points.
Artifact Mapping
Analyzing legacy spreadsheets and forms used daily to capture hidden workflow logic.
Early Validation
Testing information architecture with real clinicians before any high-fidelity work.
Nurse / Clinician
"I need to quickly understand patient context and risks, so I can make a defensible decision without missing critical details."
Ops Lead
"I need visibility into task status, so I can meet SLAs and avoid escalations before they become crises."
QA Specialist
"I need early signals of risk, so we can intervene before performance drops affect our compliance ratings."
Extraction is not Understanding
Users scan more than read. Large datasets overwhelm. Information without context is noise.
Progressive Disclosure
Show summaries first. Apply Gestalt principles to guide hierarchy. Reveal complexity on demand.
Low AI Trust
Reviewers distrust "black box" decisions. In healthcare, errors carry irreversible, personal consequences.
Explainable AI
AI suggests, humans decide. Confidence scores and source evidence always visible before any action is taken.
Compliance Anxiety
Fear of missing critical documentation during audits paralyzes reviewers mid-task.
Invisible Compliance
Built-in confirmations and error prevention. HIPAA compliance runs silently in the background — always present, never blocking.
Above the Surface — User Experience
Below the Surface — Compliance Engine
Reviewer Dashboard
Left-aligned progressive filtering by Priority, Type, and Status. High-contrast status indicators provide immediate visual hierarchy. Clear primary actions tied directly to SLA-driven workflows. Reduces cognitive load through instant workload clarity.
Document Intake and Task Management
Drag-and-drop upload with real-time file validation prevents incomplete submissions. Mandatory metadata capture enforced before any task enters the pipeline. Clear inline visual feedback on upload success and format compatibility.
AI-Assisted Medical Review
The "Examine and Update" architecture juxtaposes the source document alongside the AI decision panel — eliminating tab-switching. AI-extracted answers are visually distinct from manual inputs. Mandatory human confirmation required for all AI outputs.
Explainable AI Summary View
AI accuracy metrics and field extraction confidence levels grouped upfront before review begins. Visual indicators flag low-confidence extractions. A targeted validation tab prevents blind trust in automation results.
Role-Based — Users only see features relevant to their permissions level.
Task-Oriented — Navigation follows natural clinical workflow sequences.
Architecting Scale: A Playbook for Product Design and Governance
A comprehensive framework for scaling enterprise product design — covering governance, component architecture for personalization, and translating UX into measurable business outcomes.
Craft and Governance
Solving visual and behavioral inconsistency through centralized systems, strict semantic versioning, and continuous auditing workflows that prevent future decay.
Product Optimization
Balancing cognitive load and performance through progressive disclosure, logical hierarchy, and role-based personalization at the component architecture level.
Leadership and Impact
Translating UX into business outcomes, bridging engineering constraints, and navigating feature evolution — speaking the language of stakeholders, not just designers.
Robust product architecture — modular systems and cross-platform tokens ensure consistency.
Managerial leadership — solving complex, dynamic, role-based workflows at scale.
Governance and craft — aligning stakeholders, navigating constraints, focusing on business impact.
Discovery
- Audit components
- Identify duplicates
- Analyze spacing
Prioritization
- Usability impact
- Accessibility depth
- Tech debt triage
Restructuring
- Consolidate elements
- Standardize tokens
- Write instructions
Governance
- Define guidelines
- Build approval loops
- Track library adoption
Top Layer — Contextual Overlays
Adapts to user role, data state, permissions, and preferences. Dynamic cards and context-aware notifications that flex without breaking system logic.
Middle Layer — Shared Patterns
Component behaviors and content hierarchy rules. The behavioral contract between the design system and all product teams.
Base Layer — Tokens and Layout
Structural constraints, spacing, and layout rules. Ensures visual stability across all permutations and screen sizes.
| Dimension | Admin | Manager | End-User |
|---|---|---|---|
| Focus | Configuration and Governance | Team Oversight and Approvals | Task Completion and Productivity |
| Flow | Permission mapping, system setup | Reporting tutorials, workflow routing | Progressive feature discovery, avoiding Day-1 overload |
| Defaults | Administrative dashboards | Team metric views | Active task queues |
Progressive Disclosure
Reveal information step-by-step to reduce cognitive overload. Improves comprehension, simplifies decision-making, and increases task completion rates.
Applied to: Multi-step onboarding flows, complex form inputs, dashboard drill-downs.
The Friction Diagnostic Loop
Map the user flow to measure completion rates, exit points, time spent, and error frequency — finding the exact step where behavioral friction occurs.
Phase 1 — Identify (The Funnel)
- Map completion rates at each step
- Identify exit points and drop-off moments
- Measure time spent on each action
- Track error frequency and recovery paths
Phase 2 — Diagnose (The Loop)
- Session recordings and heatmap analysis
- User interviews and usability testing
- Evaluate cognitive load and form complexity
- Assess trust issues and performance delays
Translating UX to Business
Avoid UX jargon. Translate "cognitive overload" into "increased drop-offs and support costs." Frame all design decisions around conversion, retention, and revenue. Speak the language of decisions, not design.
Bridging Engineering and PM
Clarify constraints openly and prioritize impact vs. effort. Propose phased MVP rollouts and iterative compromises when ideal UX is not immediately feasible. The best design that ships beats the perfect design that does not.
Owning Post-Launch Failures
When a simplified workflow failed because contextual guidance was removed, we did not defend the design. We analyzed behavior data and reintroduced cues. Iteration and accountability, not defensiveness.
Treated as a Shipped Product
Semantic versioning, backward compatibility, release notes, deprecation strategies, and migration documentation.
Behavioral, Not Just Visual
Success is measured by team adoption rates and behavioral consistency — not by how beautiful the component library looks.
Conversion, Retention, Revenue
Every design decision traced to a business metric. No "nice to have" — only "what problem does this solve and how do we measure it."
VAM: Redesigning the Legacy Bank User Agent Application
End-to-end redesign of a fragmented, outdated bank user agent management platform — transforming a tool that frustrated frontline staff into a streamlined, intuitive system for managing customer accounts, policies, and transactions.
What is VAM?
A B2B SaaS platform used by bank user agents to manage customer accounts, process policy updates, handle transactions, and resolve service requests — all from a single interface used daily by frontline staff.
Platform had Decayed
The existing system was built on outdated UI patterns with no design system, inconsistent interactions across modules, and no alignment between agent workflows and the software's information architecture.
End-to-End Ownership
Led the complete UX redesign — from discovery research and journey mapping to interaction design, prototyping, and final delivery. Collaborated with product managers and engineering leads throughout.
Reduce Task Completion Time
Eliminate redundant steps and navigation dead-ends so agents can complete customer service tasks faster and with fewer errors.
Support All Workflows in One App
Consolidate account management, policy handling, and transaction processing into a single, coherent information architecture.
Build a Scalable Design System
Establish reusable components and design patterns that the engineering team can extend as new features are added to the platform.
Discover
Contextual interviews with bank agents, observation of daily workflows, competitive analysis of banking platforms.
Define
Persona creation, journey mapping, pain point synthesis, KPI definition and project goal alignment.
Develop
Information architecture, wireframing, user flows, component library creation, and iterative prototyping.
Deliver
High-fidelity mockups, design system handoff, usability testing, dev collaboration, and shipped product.
Contextual Interviews
Conducted structured interviews with frontline bank agents across multiple branches. Focused on task frequency, pain points with current software, workarounds they had developed, and unmet needs that caused daily friction.
Workflow Observation
Observed agents in live customer service scenarios. Mapped the actual steps taken versus the system's intended flow — surfacing navigation dead-ends, redundant screens, and missing feedback states.
Navigation Overload
Agents navigated an average of 7 screens to complete tasks that should require 2–3. The menu structure was organised by system logic, not agent workflow.
No System Feedback
The legacy system provided no confirmation states, loading indicators, or error explanations. Agents frequently repeated actions unsure whether they had succeeded.
User Efficiency Blockers
Inconsistent table layouts, non-standard date pickers, and unlabelled icon buttons created a steep learning curve even for experienced agents.
Slow, Fragmented Workflows
Policy updates, account lookups, and transaction processing each lived in separate modules with no cross-linking — forcing agents to memorise navigation paths and lose context when switching tasks.
Agents Did Not Trust the System
Without success confirmations or inline validation, agents regularly escalated to supervisors to verify actions — creating bottlenecks and undermining confidence in the tool entirely.
Priya S.
"I know what I need to do, but the system makes me take 10 steps to do what should take 2. Every delay frustrates the customer waiting in front of me."
Ravi M.
"I can barely navigate the system on my own. I'm always asking colleagues where to find things — it's embarrassing in front of customers."
Ananya P.
"I spend too much time verifying actions my team should be able to handle independently. The system gives them no confidence."
| Stage | Login | Search Customer | View Account | Update Policy | Confirm Action |
|---|---|---|---|---|---|
| Actions | Enter credentials, navigate to module | Use search bar across 3 possible entry points | Parse dense table to locate account | Navigate 5 sub-screens to locate field | No confirmation shown — repeat action |
| Pain Level | Low | Medium | High | Critical | Critical |
| Feeling | Neutral | Mildly frustrated | Confused, scanning | Frustrated, lost | Anxious, uncertain |
Reduce Average Task Completion
Target: Cut average steps per task from 7 to 3. Measured through usability testing session recordings and time-on-task analysis.
Increase Agent Confidence Score
Target: Improve self-reported confidence with the system from 2.1/5 to above 4/5. Measured through post-session SUS surveys.
Reduce Supervisor Escalations
Target: Reduce task-related escalations to supervisors by 60%. Measured through operational logs before and after launch.
Feature-grouped navigation
System organised by module logic, not agent workflow — forcing mental mapping at every step.
Task-oriented navigation
Restructured IA around the most frequent agent tasks. Primary nav reflects jobs-to-be-done, not system modules.
No system feedback or confirmation
Agents had no way to know if their actions succeeded, pending, or failed.
Persistent status visibility
Every action now produces an inline status response. Toast notifications, confirmation dialogs, and real-time form validation built into every flow.
Steep learning curve for new agents
Unlabelled icons, inconsistent patterns, and zero onboarding created a hostile first-use experience.
Progressive disclosure and labelled UI
All interactive elements labelled. Contextual help text added to complex fields. Complex tasks broken into guided, stepped flows.
Primary Navigation — Task-Oriented
Dashboard, Customer Search, Account Management, Policy Updates, Transactions, Reports. Organised by agent job-to-be-done, not system modules.
Contextual Actions — Inline and Accessible
Quick actions surfaced directly on customer records. No more navigating to a separate module to update a policy or process a transaction.
Feedback Layer — Always Visible
System status always communicated. Success confirmations, error explanations, loading states, and validation messages present at every decision point.
Low-Fidelity Iteration
Produced rapid lo-fi wireframes for all core flows: customer search and account view, policy update flow, transaction processing, and the dashboard summary. Tested with agents before investing in high-fidelity.
Usability Sessions with Agents
Ran 3 rounds of usability testing with 5 bank agents each. Used task-based scenarios reflecting real work conditions. Iterated on navigation labels, action placement, and form field ordering based on findings.
Agent Dashboard
At-a-glance overview of pending tasks, recent customer interactions, performance metrics (daily transactions processed, open cases), and quick-access shortcuts to the most frequent workflows.
Customer Account View
Consolidated single-page view of all customer data — account summary, policy status, transaction history, and open cases — eliminating the need to navigate between modules for a complete customer picture.
Policy Update Flow
Stepped guided flow reduced from 7 screens to 3. Each step shows only relevant fields, with inline validation, clear progress indication, and a review-and-confirm final step before any change is submitted.
Transaction Management
Redesigned transaction table with advanced filtering, sortable columns, status indicators, and bulk action support. Confirmation dialogs added for all irreversible operations.
How I Work
Discover
- Stakeholder interviews
- Contextual shadowing
- Artifact analysis
- Competitive audits
Define
- Synthesis & insights
- Persona creation
- Journey mapping
- KPI definition
Design
- IA & wireframing
- High-fidelity mockups
- Design system tokens
- Interactive prototyping
Validate
- Usability testing
- Clinical walkthroughs
- Friction diagnostics
- Design QA audits
Deliver
- Developer handoff
- Implementation syncs
- Impact monitoring
- System governance
Let's work together.
Open to full-time UX and product design roles, contract engagements, and consulting for complex design challenges. Especially interested in healthcare, enterprise, and AI products where the stakes are high and clarity matters most.
I work best with teams that value research, move with intention, and care about the humans using what they build.
Open to Work
Looking for a designer who brings both craft and strategic thinking — who can navigate regulated domains, complex systems, and human-critical products without losing sight of the user.