CASE STUDY

Checklist AI / Verification Assistant

Designing a compliance assistant where AI extraction accelerates review, while humans keep final control over verification outcomes.

RoleProduct Designer
FocusVerification UX
ToolsFigma
Project visual coming soon screen

Project Context

Compliance workflows needed speed and traceability

The product supports teams that verify documents against strict instructions and policy rules. The challenge was speeding up analysis while keeping review decisions auditable.

The UX Challenge

Make verification clear from upload to report

  • Guide checklist instructions before analysis begins
  • Support PDF upload and extraction review without confusion
  • Surface issues by severity and evidence context
  • Preserve a clear human review and approval step
  • Produce reports that are clear for audits and handoff

My Role

My Role

I designed the end-to-end verification flow: instruction setup, extraction review, issue resolution, and final reporting for operational teams.

Interaction Model / Core Flow

A structured path from document to decision

1. Checklist setup

Users define criteria and verification scope before uploads.

2. Document upload

PDF and supporting files are added with clear processing feedback.

3. AI extraction review

Extracted fields are shown with confidence cues and source references.

4. Issue detection

Potential mismatches and missing items are grouped by priority.

5. Human verification and report

Reviewers confirm outcomes and generate a clear, traceable report.

Image placeholder — add checklist, upload, extraction, and issue workflow

Key UX Decisions

Decisions for trust and review confidence

Instruction clarity

Checklist criteria remain visible throughout the process.

Source transparency

Each issue links back to extracted document evidence.

Severity framing

Issue hierarchy helps teams prioritize what needs review first.

Approval safeguards

Final verification always requires explicit human confirmation.

Reporting readability

Outputs are structured for quick team sharing and audit use.

Design System / Interface Logic

Consistent patterns for compliance-critical tasks

  • Reusable instruction, upload, and validation components
  • Unified visual language for warnings, flags, and confirmations
  • Clear states for extracted, pending, and verified data
  • Stable spacing and table patterns for issue management
  • Consistent report layout for trust and traceability
Project visual coming soon layout
Focus area: trust and clarity in AI-supported verification

Final Solution

A verification assistant that supports reviewers

The final product combines checklist logic, AI extraction, and issue guidance in one coherent workflow where human judgement remains central.

Compliance checks became easier to review, explain, and hand off.

Outcome

More structured review with clearer reporting

The interface reduced confusion during complex verification tasks, and implementation moved faster because states and steps were clearly defined.