🏠 Home
Modules
📘 Module 01 📘 Module 02 📘 Module 03 📘 Module 04 📘 Module 05 📘 Module 06 📘 Module 07 📘 Module 08 📘 Module 09 📘 Module 10 📘 Module 11 📘 Module 12 ❓ FAQ 🚪 Sign Out
Module 10 of 12  ·  Hello Pharma AI Upskilling Program

AI Collaboration & Workflow Design

Design AI as a team member, not a tool.

📋 Official Content ⏱️ ~25 min read ✏️ Exercise included
📚 5 sections
🏢 Hello Pharma
🎯 5 objectives
⚕️

Hello Pharma AI Upskilling Program

This module is part of Hello Pharma's internal AI capability-building programme, designed to help every team member work with AI professionally and responsibly.

Official Content

🎯 Learning Objectives

  • Define AI roles within professional workflows
  • Use AI as a research assistant, analyst, and editor
  • Design adversarial AI roles for quality improvement
  • Build multi-persona prompting systems
  • Create team AI workflows that scale

1. Defining AI Roles in Workflows

AI RoleBest ForKey Constraint
Research AssistantGathering, summarising informationAlways verify facts independently
AnalystInterpreting data, identifying patternsProvide context and decision framing
EditorImproving clarity, structure, and toneDefine editorial standards explicitly
Devil's AdvocateChallenging assumptions, finding weaknessesAssign adversarial role explicitly
DrafterProducing first-draft content for human refinementExpect and plan for editing
SimulatorModelling stakeholder reactionsDefine the stakeholder precisely

2. AI as Analyst

You are a senior business analyst. Context: [business situation and decision being faced] Data: [data or description] Tasks: 1. Identify the 3 most important patterns in this data 2. What does each pattern suggest about the underlying business? 3. Which pattern most strongly influences the decision at hand? 4. What additional data would most change your analysis? Interpret, don't describe. Connect every finding to the business decision.

3. AI as Editor

You are a senior editor at a business publication. Edit the following content against these standards: 1. Every paragraph opens with its main point 2. No sentence over 20 words 3. Remove any sentence that doesn't add new information 4. Flag (but don't fix) any claim that needs a source 5. Preserve the author's voice — do not rewrite for style [Paste content] Show edits: [CHANGED: original → revised] inline. Provide a summary of the 3 biggest structural issues at the end.

4. AI as Devil's Advocate & Challenger

You are a sceptical board member who has seen many strategies fail. Review the following plan: [plan] 1. Challenge the core assumption underlying this strategy 2. Identify the 3 most likely failure modes 3. Ask the 5 questions the team hasn't answered yet 4. Name the one scenario that would completely invalidate this plan Be direct. Do not soften the critique.
💡 Pre-Mortem Technique

"Imagine it is 12 months from now and this plan has failed completely. What went wrong? Describe the most likely failure story in specific detail."

5. Multi-Persona Prompting

Simulate a meeting between three stakeholders reviewing: [proposal] Stakeholders: - CFO (cost, ROI, financial risk) - Head of Compliance (regulatory exposure) - Operations Director (implementation feasibility) For each stakeholder: 1. Their primary concern (1 sentence) 2. Their key question to the proposer (1 sentence) 3. Vote: Support / Conditional / Oppose, and why Conclude: The most likely consensus outcome of this meeting.

✏️ Module 10 Exercise

Map one of your current workflows and identify where AI could play a bounded role. Define the role precisely, write the prompt, and run a test. Then apply the devil's advocate prompt to something you're about to present or publish.

🔑 Key Takeaways

  • Defining a bounded AI role produces better outputs than open-ended tasking
  • The analyst role requires explicit decision framing — connect data to choices
  • The editor role requires explicit editorial standards, not generic improvement
  • Devil's advocate prompting surfaces weaknesses before they become expensive
  • Multi-persona prompting simulates stakeholder perspectives efficiently
  • Human judgement remains essential — AI roles should augment, not replace, critical thinking