PR with User Workflow Impact

Generate pull requests that clearly communicate how code changes affect real users. Transforms technical diffs into concrete before/after examples with emotional resonance.

When to Use

• Creating a pull request
• Writing commit messages for significant changes
• Describing the impact of a feature, fix, or refactor
• Reviewing changes before merge

Core Principle

Every change affects a user doing something. Document the "who does what" and "how it changes for them" with concrete examples, not abstract descriptions.

PR Template

Every PR includes this section:

## User Workflow Impact

### [Workflow Name]
**User action:** [What the user is trying to accomplish]

| Before | After |
|--------|-------|
| [Concrete example of the problem with realistic details] | [Concrete example of the fix with realistic details] |

**Impact:** [Emotional language + business outcome]

Writing Effective Before/After Tables

WRONG - Abstract descriptions

RIGHT - Concrete and specific

Writing Effective Impact Statements

WRONG - Clinical/factual only

Page loads faster now.

Users can now see more precise categories.

Error handling is improved.

RIGHT - Emotional + business outcome

Eliminates the "is it broken?" moment. Users no longer abandon the dashboard before it loads.

Patterns that were buried in misleading labels now surface on their own. Users act on real insights, not aggregation artifacts.

Developers stop guessing what went wrong. The error tells them exactly what to fix.

Emotional Language Bank

Use these phrases to make pain/relief tangible:

Pain words:

• "frustrating"
• "buried"
• "hidden"
• "misleading"
• "forced to manually"
• "wasted time"
• "mysterious"
• "silently fails"
• "guessing"

Relief words:

• "eliminates"
• "surfaces automatically"
• "immediately visible"
• "no longer need to"
• "obvious"
• "tells them exactly"
• "one click"

Discovery framing:

• "X was hidden in Y → X now surfaces as its own insight"
• "What was buried in [broad category] now appears as [specific item]"

Examples by Change Type

Performance improvement

### Dashboard Loading
**User action:** Opening the analytics dashboard to check daily metrics

| Before | After |
|--------|-------|
| 4.2 second blank screen—users often refresh thinking the page is frozen | 0.8 second load with skeleton UI appearing immediately |

**Impact:** Eliminates the "is it broken?" moment. Users start their day with data instead of frustration.

Bug fix

### Form Submission
**User action:** Submitting the contact form

| Before | After |
|--------|-------|
| Entering "José García" causes silent failure—users retry multiple times before giving up or removing accents | "José García" submits successfully. All UTF-8 characters work. |

**Impact:** Users with accented names no longer have to "anglicize" their input to get through.

Feature addition

### Data Export
**User action:** Exporting results for a stakeholder presentation

| Before | After |
|--------|-------|
| Only CSV available—users manually convert to Excel, recreating charts and formatting | Direct Excel export with formatted tables and charts. One click to presentation-ready. |

**Impact:** Eliminates the CSV-to-Excel conversion ritual. "Analysis complete" to "ready to present" in seconds.

Refactor/API change

### API Error Handling
**User action:** Debugging a failed API integration

| Before | After |
|--------|-------|
| `{"error": "Something went wrong"}` | `{"error": "INVALID_DATE_FORMAT", "field": "start_date", "expected": "ISO 8601", "received": "12/25/2024"}` |

**Impact:** Developers stop guessing what went wrong. The error tells them exactly what to fix.

Algorithm/data change

### Category Review
**User action:** Reviewing AI-generated categories to find actionable patterns

| Before | After |
|--------|-------|
| A category "Cancel and Refund Requests" with 47 docs—but 38 only mention cancellation, hiding that pricing is the real issue | Two categories: "Cancellation Due to Pricing" (38 docs) and "Refund Delays" (6 docs). The primary driver is obvious. |

**Impact:** Patterns that were buried in misleading labels now surface on their own. Users act on real insights, not aggregation artifacts.

Security/infrastructure change

### Session Management
**User action:** Working on a long document without saving

| Before | After |
|--------|-------|
| Session expires silently after 30 minutes—users lose unsaved work with no warning | 5-minute warning before expiration with one-click extend. Auto-save every 60 seconds. |

**Impact:** No more lost work from surprise logouts. Users can focus on their task without watching the clock.

Key Rules

1. Never use abstract comparisons - "better," "improved," "faster" need concrete proof
2. Include realistic details - times, counts, error messages, names, category names
3. Testable claims - someone could verify the before/after by using the product
4. No technical jargon - never mention parameters, thresholds, implementation details, or internal code concepts
5. One scenario per section - if multiple impacts, use multiple workflow sections
6. Business outcomes - connect to why stakeholders should care, not just what changed

How to Identify User Workflows

When analyzing a diff, ask:

1. Who uses this? (End users, developers, admins, mobile users, etc.)
2. What are they trying to accomplish? (Login, search, export, debug, etc.)
3. What was painful before? (Slow, confusing, broken, manual, hidden)
4. What's better now? (Fast, clear, working, automatic, visible)
5. Why should someone care? (Time saved, errors prevented, decisions enabled)

Anti-Patterns to Avoid

Too technical

• "Refactored useEffect dependencies"
• "Updated webpack config"
• "Changed clustering scale from 0.7 to 0.5"
• "Fixed race condition in state management"

Too vague

• "Improved performance"
• "Fixed bugs"
• "Better UX"
• "More accurate results"

No user context

• "Added caching layer"
• "Migrated to new API"
• "Updated dependencies"

Multiple User Types

When changes affect different user groups, include separate sections or a summary:

## User Workflow Impact

### For End Users: Dashboard Loading
**User action:** Opening the app to check their data
| Before | After |
|--------|-------|
| 4s blank screen | 0.8s with skeleton UI |
**Impact:** Eliminates the "is it broken?" moment.

### For Developers: Error Debugging
**User action:** Investigating failed API calls
| Before | After |
|--------|-------|
| Generic error message | Structured error with field and expected format |
**Impact:** Debugging time drops from minutes to seconds.

Quick Checklist

Before finalizing a PR, verify:

• Identified which users are affected
• Described what they're trying to do (the workflow)
• Showed concrete before/after with realistic details
• Used emotional language in impact statement
• Connected to business outcome (why it matters)
• Avoided technical jargon and implementation details
• Claims are testable by using the product