Skip to content

persona-first-decision-facilitation-loop.md

Latest commit

 

History

History
240 lines (177 loc) · 7.47 KB

persona-first-decision-facilitation-loop.md

File metadata and controls

240 lines (177 loc) · 7.47 KB
Raw

Persona-First Decision Facilitation Loop (PDF Loop)

What Is The PDF Loop?

The Persona-First Decision Facilitation Loop (PDF Loop) is a conversational pattern for product-management prompts.

It is designed to feel like expert guidance, not a chatbot quiz:

  • One targeted question at a time
  • Three context-aware options at decision points
  • Recommended option first with rationale
  • Persona language first, business translation second
  • Fast adaptation after each answer
  • Decision summary with assumptions and next steps

Why This Exists

Most prompt experiences fail in one of two ways:

  1. Unstructured drift: "tell me more..." without decision progress
  2. Form overload: long questionnaires that push setup labor onto users

The PDF Loop sits in the middle:

  • Human and conversational
  • Programmatic and reliable
  • Decision-forward and pedagogic

Non-Negotiable Rules

  1. Workload Inversion The assistant must do heavy lifting. Ask for minimum viable context, then propose likely scopes/options.

  2. One-Question Turns Ask one targeted question. Do not batch multiple questions unless explicitly requested.

  3. Decision Fork Shape At each meaningful fork, offer exactly 3 options:

  • Option 1 recommended first
  • Each option includes short rationale
  • User can respond with 1, 2, 3, 1 and 3, or custom direction

  1. Persona-First Wording Frame recommendations in user/persona language first. Add one-line business translation second when useful.

  2. Progress Visibility Show momentum with concise progress updates (example: Progress: 2/4 core inputs captured).

  3. Decision Closure Summarize what was decided, why, what happens next, and assumptions to validate.

Where To Use It

Use PDF Loop by default in:

  • /prompt-generators assets
  • Decision-heavy /storytelling assets
  • Any prompt that requires tradeoffs and sequencing

Do not force PDF Loop when:

  • User requests one-shot output
  • Task is pure transformation (rewrite, summarize, format)
  • Creative ideation is intentionally unconstrained

Session Lifecycle

Step 0: Set Expectations

In one short message:

  • Goal of the session
  • Approximate time
  • What the flow will look like

Step 1: Capture Minimum Context

Ask for smallest set of inputs needed to infer structure.

Good:

  • "Who is the customer and what painful moment are they in?"

Bad:

  • "Define your full journey map stages, touchpoints, and KPIs."

Step 2: Propose Scope Options

Assistant proposes 3 candidate scopes based on context.

Step 3: Run Decision Forks

For each major decision:

  • 3 options
  • one recommended first
  • persona-first language
  • optional business translation

Step 4: Generate Artifact

Produce requested artifact in structured format.

Step 5: Close The Loop

Include:

  • Decisions made
  • Assumptions to validate
  • 3 (or more) concrete next-step options

Canonical Decision-Turn Template

Based on what you shared, here are the three best paths:

1. [Persona-first option] (Recommended) - [why now]
   Business translation: [optional]
2. [Persona-first option] - [tradeoff]
   Business translation: [optional]
3. [Persona-first option] - [tradeoff]
   Business translation: [optional]

Reply with `1`, `2`, `3`, `1 and 3`, or tell me your own path.

Option Quality Rubric

Every option set should score well on:

  • Context fit: grounded in what user already said
  • Mutual exclusivity: distinct enough to choose between
  • Actionability: clear implication if chosen
  • Tradeoff clarity: users can see what they gain/lose
  • Language fit: understandable to persona, not just internal teams

Anti-Patterns (And Fixes)

Anti-Pattern: Burden-Shifting Question

"What journey are we mapping?"

Fix: "Who is the customer, and what painful moment are they in now?" Then assistant proposes 3 candidate journey scopes.

Anti-Pattern: Jargon-First Options

"Increase qualified purchase intent."

Fix: "Help us confidently pick 2-3 options we can both afford." Business translation can follow.

Anti-Pattern: No Closure

Artifact generated, no decision summary, no assumptions, no next steps.

Fix: Always end with Decisions + Assumptions + Next-step options.

Worked Example 1: Customer Journey Mapping

Input (minimal)

"Fred, father of 3, overwhelmed by SUV/minivan choices and affordability."

Assistant Proposal

  1. Discovery to shortlist (Recommended)
  2. Budget and financing confidence
  3. Household decision alignment

Persona-First Outcome Option Example

  1. "Help us pick 2-3 safe options we both can afford." (Recommended) Business translation: increase qualified purchase intent.

Worked Example 2: Positioning

Input (minimal)

"Ops managers are buried in outage noise and cannot prioritize incidents."

Assistant Proposal

  1. Position around decision confidence under pressure (Recommended)
  2. Position around reduced alert volume
  3. Position around collaboration visibility

Persona-First Differentiation Example

  1. "Know what to fix first in minutes, not hours." (Recommended) Business translation: faster time-to-value and lower churn risk.

Worked Example 3: Proto-Persona

Input (minimal)

"First-time team lead trying to run weekly 1:1s effectively."

Assistant Proposal

  1. Coaching-first persona lens (Recommended)
  2. Process-compliance lens
  3. Career-growth lens

Persona-First Hypothesis Example

  1. "I want to support my team without sounding scripted." (Recommended) Business translation: higher manager confidence and retention.

Implementation Checklist For Prompt Authors

Before publishing, verify:

  1. Prompt asks one targeted question at a time.
  2. Prompt uses workload inversion (assistant proposes structure).
  3. Decision points use exactly 3 context-aware options.
  4. Recommended option is first and justified.
  5. Options are persona-first, with optional business translation.
  6. User can answer with numeric choice, combinations, or custom direction.
  7. Prompt includes progress signals.
  8. Prompt closes with decision summary + assumptions + next steps.
  9. Comment block explains why this structure works.

Copy/Paste Starter Block For New Prompts

### Facilitation Rules
1. Ask one targeted question at a time.
2. Apply workload inversion: gather minimal context first, then propose likely scopes/options.
3. At decision points, offer exactly 3 options, with one recommended first.
4. Phrase options in persona language first; add business translation second.
5. Accept `1`, `2`, `3`, `1 and 3`, or custom direction.
6. Show progress and close with decisions, assumptions, and next actions.

Naming

When referencing this method, use:

  • Persona-First Decision Facilitation Loop
  • Short form: PDF Loop