Compliance teams can't evaluate their own workflows for automation potential.
Evaluating automation potential requires both deep process knowledge (what actually happens, what can go wrong, what regulators expect) and technical understanding (what's feasible, what tools exist, what the ROI looks like). That combination is rare. Compliance professionals lack the technical lens; engineering teams lack the regulatory context.
Current workarounds: gut instinct, expensive consultants, or nothing at all. Most compliance teams don't systematically evaluate their workflows for automation potential. They just keep doing things the way they've always done them.
What I prioritized, what I cut, and why
A 3-week build forces ruthless prioritization. Here's how I scoped:
Pre-built workflow templates
10 example workflows from real compliance scenarios. Templates lower friction and demonstrate domain expertise to hiring managers.
Workflow visualization
Process flowcharts would add complexity without proving the core value prop. The AI analysis is the differentiator, not diagrams.
Categorized AI output
Results organized into Quick Wins, High Impact, Process Improvement, Data Quality, and Strategic Value. Mirrors how teams actually prioritize.
Cost-benefit calculator
Nice-to-have for production, but not essential for demonstrating product thinking. Users can do their own math.
Professional PDF export
Export functionality signals "real tool" rather than "demo toy." Compliance teams need to share recommendations with stakeholders.
Multi-workflow comparison
Single-workflow analysis proves the concept. Comparison features can come later once the core experience is validated.
Select a template or describe your workflow, get prioritized recommendations.
Users pick from 10 pre-built compliance workflow templates (lobbying disclosure, insider trading pre-clearance, SAR filing review, etc.) or describe their own process. They add time estimates, click analyze, and receive AI-generated automation opportunities organized by category and complexity.
Results display in a carousel with color-coded cards. Users can export the analysis as a PDF to share with leadership or include in a business case.
Technical and product insights
Serverless functions solve the CORS problem
My first attempt called Claude directly from the frontend. CORS blocked it immediately. Vercel serverless functions handle API calls server-side, keeping credentials secure and eliminating cross-origin issues entirely.
Native browser printing beats PDF libraries
Instead of jsPDF, I used native browser printing with @media print CSS. The trick: print-color-adjust: exact preserves background colors. Cleaner code, no dependencies, better output.
Scope discipline is the hardest part
I score at the 99th percentile for Idea Productivity, which means I generate ideas constantly. The solution: strict timeboxes and a rule that no new features ship until the current version is live.
Design system consistency builds trust
Keeping every page consistent (same typography, same colors, same spacing) made the whole portfolio feel intentional. Inconsistency signals carelessness.
What's next
If I continued development, these are the natural extensions: