PRD Writer Skill

Specialized knowledge for generating structured PRDs that align with AskElephant's product vision.

When to Use

Creating new PRDs
Updating existing requirements
Reviewing PRD drafts
Defining feature scope

Before Writing: Strategic Alignment Check

CRITICAL: Before drafting ANY PRD:

Read pm-workspace-docs/company-context/product-vision.md
Read pm-workspace-docs/company-context/strategic-guardrails.md
Check pm-workspace-docs/company-context/personas.md
Look for related PRDs in pm-workspace-docs/initiatives/

Pushback Triggers

Stop and ask clarifying questions if you detect:

🚩 Unclear Outcomes

"What business outcome does this enable?"
"How would we know if this is working?"

🚩 Missing Persona Context

"Which persona needs this—reps, leaders, CSMs, or RevOps?"
"What's their current pain?"

🚩 No Evidence

"What evidence suggests customers need this?"
"If we didn't build this, what happens?"

🚩 Revenue Impact Unclear

"How does this help land, expand, or retain?"
"Does this help win deals we're losing?"

🚩 Trust/Privacy Implications

"How does this interact with privacy?"
"What happens when this fails?"

🚩 Anti-Vision Warning

"This sounds like generic AI notes—how is this uniquely AskElephant?"

🚩 Experience Gap

"Where's the discovery plan? How do users find this?"
"What happens on day 2? What brings them back?"
"How will we collect feedback after launch?"
"Can a customer enable this without someone on our side hand-holding them?"

Outcome Chain Requirement

Every PRD MUST include:

[Feature] enables [user action]
  → so that [immediate benefit]
    → so that [behavior change]
      → so that [business outcome]

If the chain cannot be articulated, do not proceed with PRD. Recommend discovery.

PRD Structure

# [Feature Name] PRD

## Overview

- **Owner:**
- **Target Release:**
- **Status:** Draft
- **Strategic Pillar:** [Trust / Data Knowledge / Trend Visibility]

## Outcome Chain

[Feature] enables [action]
→ so that [benefit]
→ so that [behavior change]
→ so that [business outcome]

## Problem Statement

What problem? Who has it? Why now? What evidence?

### Evidence

- [ ] User quotes
- [ ] Churn/support data
- [ ] Competitive pressure

## Goals & Non-Goals

### Goals (Measurable)

- [ ] Goal with success metric

### Non-Goals

- Explicit exclusion with reasoning

## User Personas

### Primary: [Name]

- **Job-to-be-done:**
- **Current pain:**
- **Success looks like:**
- **Trust factors:**

### Secondary: [Name]

- **Job-to-be-done:**
- **Current pain:**
- **Success looks like:**
- **Trust factors:**

### Tertiary: [Name] (if applicable)

- **Job-to-be-done:**
- **Current pain:**
- **Success looks like:**
- **Trust factors:**

## User Stories (Per Persona - REQUIRED)

**CRITICAL:** Create explicit user stories for EACH persona defined above. These feed into FigJam diagram generation.

### Primary Persona Stories

- As a [Primary Persona], I want [action] so that [benefit]
- As a [Primary Persona], I want [action] so that [benefit]

### Secondary Persona Stories

- As a [Secondary Persona], I want [action] so that [benefit]

### Tertiary Persona Stories (if applicable)

- As a [Tertiary Persona], I want [action] so that [benefit]

## Shared Customer Journey

**REQUIRED:** Define the shared journey that applies across personas.

### Current State (Pain Points)

1. Step 1 → Pain point
2. Step 2 → Pain point
3. Step 3 → Pain point

### Future State (With Feature)

1. Step 1 → Improvement
2. Step 2 → Improvement
3. Step 3 → Improvement

### Transformation Moment

What is the key "aha moment" when users realize the value?

## End-to-End Experience Design (REQUIRED)

**CRITICAL:** Every feature must describe the full user journey, not just the feature itself. (Rob Henderson framework)

### 1. Discovery -- How does the customer know this exists?

<!-- How will users find this feature? In-app announcement, email, onboarding, organic discovery? -->

### 2. Activation -- How do they enable/configure without hand-holding?

<!-- What does the zero-config or first-time setup look like? Can they do it alone? -->

### 3. Usage -- What does the first interaction look like?

<!-- Walk through the first time a user actually uses this. Be specific. -->

### 4. Ongoing Value -- What value do they get on day 2, week 2, month 2?

<!-- What brings them back? What compounds over time? -->

### 5. Feedback Loop -- How do we know if this is working for them?

<!-- In-app survey, NPS, usage analytics, support ticket patterns? Be specific. -->

## Requirements

### Must Have (MVP)

### Should Have

### Could Have

## User Flows

### Flow: [Name]

**Trigger:**
**Steps:** 1 → 2 → 3
**Outcome:**
**Error states:**
**Trust recovery:**

## Trust & Privacy Considerations

## Success Metrics

- **North star:**
- **Leading indicators:**
- **Guardrails:**

## Strategic Alignment

- [ ] Outcome chain complete
- [ ] Persona validated
- [ ] Trust implications assessed
- [ ] Not in anti-vision territory
- [ ] End-to-end experience: All 5 steps addressed (Discovery, Activation, Usage, Ongoing Value, Feedback)
- [ ] Feedback method defined for post-launch learning
- [ ] Ownership assigned (PM, Eng Lead, Design Lead)

## Launch Materials Needed

- [ ] Revenue team training deck
- [ ] Help center article
- [ ] Changelog entry
- [ ] In-app announcement / tooltip
- [ ] Slack #product-updates post
- [ ] Customer communication (if applicable)

## Open Questions

Save Location

pm-workspace-docs/initiatives/[kebab-case-name]/prd.md

When to Refuse

If initiative fails strategic guardrails:

List gaps/concerns clearly
Recommend specific discovery work
Suggest questions to answer first
Offer to help with discovery instead

When to Refuse: Experience Gaps

If the PRD cannot articulate all 5 experience steps, do NOT proceed. Instead:

Flag which steps are missing
Recommend specific discovery work for each gap
Suggest: "Before writing the PRD, we need to answer: How will users discover this? What does day 2 look like?"

This aligns with Rob Henderson's feedback: "I can't name a certain thing we've ever launched that actually has all five of these things." Every PRD must change that.

FigJam Customer Story Generation (REQUIRED)

After completing the PRD, generate a FigJam diagram that visualizes the customer story.

When to Generate

After PRD is complete with user stories and shared journey
Before prototype begins (can be done during /pm command)

FigJam Content Requirements

The diagram should include:

Current State (Pain) - The painful user journey with key quotes
User Stories by Persona - Each persona's primary user story
Future State (Solution) - The improved journey with the feature
Outcome Chain - Business impact flow

MCP Tool Call

Use the Figma MCP generate_diagram tool:

Server: figma
Tool: generate_diagram
Arguments:
  - name: "[Initiative Name] - Customer Story"
  - mermaidSyntax: [Generated flowchart - see template below]
  - userIntent: "Visualize the customer problem and user stories for [initiative]"

Mermaid Template

flowchart LR
    subgraph Current["CURRENT STATE"]
        direction TB
        C1["Step 1: Pain point"] --> C2["Step 2: Pain point"]
        C2 --> C3["Step 3: Pain point"]
        Pain1(["Key quote from research"])
        Pain2(["Quantified impact"])
    end
    subgraph Stories["USER STORIES"]
        direction TB
        P1["Primary: As a [persona]..."]
        P2["Secondary: As a [persona]..."]
        P3["Shared: Transformation moment"]
    end
    subgraph Future["FUTURE STATE"]
        direction TB
        F1["Step 1: Improvement"] --> F2["Step 2: Improvement"]
        F2 --> F3["Step 3: Outcome"]
        Win1(["Success metric target"])
    end
    Current -->|"Transformation"| Stories
    Stories -->|"Enables"| Future
    style Current fill:#ffcccc
    style Stories fill:#fff3cd
    style Future fill:#d4edda

Save FigJam URL

After generating, save the FigJam URL to _meta.json:

{
  "figjam_url": "https://www.figma.com/...",
  "figjam_generated": "2026-01-28"
}

This URL will be included in prototype notifications sent via Slack.

prd-writer