Restricted Access

Enter password to continue

The Decision Journey
→ Press space or right arrow to begin
The Decision Journey

Six templates. One case. The right AI tool at each step.

6 templates · ~2 hours · BeautyCo case throughout

How this deck works

Each template follows the same rhythm: Learn, review the BeautyCo evidence, use a copyable prompt, open the right AI tool, challenge the output, save your work. Discover → Diagnose → Decide → Deliver.

The Decision Challenge

"Why is increased traffic not translating into stronger business outcomes, and what should BeautyCo do next?"

"By the end of this deck, you'll have a complete BeautyCo decision workbook — ready to review at the BeautyCo Case Hub."
Discover · Template 1

Decision Framing

"What decision are we making?"
~15 minChatGPT / Gemini
🧠 Expert Perspective (RGCTO) · Ask Before Answer
Learn

Clarify the decision before analysing or recommending. Use AI as a thinking partner to ask clarification questions, formulate the decision question, identify objectives, expose assumptions, and define constraints.

❌ Common mistakeJumping straight to "improve staffing and inventory" before the decision question is even framed.
✅ Do this insteadLet AI ask up to 3 clarification questions first — a good frame surfaces assumptions before any recommendation.
Review the evidence
KPI20242025
Footfall220,000282,000
Revenue$4.8M$5.0M
Conversion24%18%
CSAT8271
Repeat purchase38%29%
CEO concern: "Traffic is increasing strongly, but business outcomes are not."
Use the template
Role: You are an experienced strategy and decision-intelligence advisor.
Goal: Clarify the decision that needs to be made.
Context: BeautyCo is a premium beauty retailer. Traffic has increased,
but revenue is almost flat. Conversion, CSAT and repeat purchase have declined.
Before answering: Ask up to three clarification questions if the decision is unclear.
Tasks: 1. Frame the decision question. 2. Identify the business objective.
3. Identify key symptoms. 4. Surface initial assumptions.
5. Identify information gaps. 6. Identify decision constraints.
Output: A Decision Framing Brief — decision question, objective, symptoms,
assumptions, information gaps, constraints.
Challenge the output
  • What did AI assume too quickly?
  • Did it frame a decision or jump to a solution?
  • Which stakeholder is missing?
Reflect + Save

"What assumption did AI make too quickly?"

🗳️ Nearpod
Save your Decision Framing Brief.
Discover · Template 2

Situation Assessment

"What is happening?"
~25 minNotebookLM + Colab + ChatGPT/Gemini
🧠 Structured Analysis (Chain-of-Thought)
Learn

Build a complete picture before diagnosing. This is the first template where tools differentiate — NotebookLM grounds you in evidence, Colab analyses the numbers, ChatGPT/Gemini synthesizes it all.

📓 NotebookLM — grounded sources+ 📊 Colab — numerical patterns→ 💬 ChatGPT/Gemini — synthesis
Optional tool layer
Evidence needTool / Agent
Internal documentsNotebookLM / Projects
External market evidenceDeep Research / Research Briefing Agent
Data patternsData Insight Agent
1
Review grounded sources
Based only on the supplied sources:
1. What direct evidence shows a conversion problem?
2. Separate observations from interpretations.
3. What customer complaints appear repeatedly?
4. What external trends matter to the decision?
5. What important evidence is still missing?
2
Analyse numerical evidence

Use the prepared BeautyCo Colab notebook — KPI trends, store comparisons, traffic vs. revenue, stock-outs vs. revenue, staffing vs. conversion.

3
Synthesize findings
Assess the situation. Organise into: 1. Internal signals
2. External signals 3. Customer/stakeholder signals
4. Emerging themes 5. Risks and opportunities 6. Evidence gaps
For every signal, separate: Observation / Interpretation / Decision implication.
Do not identify final root causes yet.
Save your Situation Assessment, then continue to Diagnose.
Diagnose · Template 3

Root Cause Analysis

"Why might it be happening?"
~20 minChatGPT/Gemini + NotebookLM + Colab
🧠 Multiple Explanations (Tree-of-Thought)
Learn

Generate competing explanations before choosing one. Expand the hypothesis space across customer, product, people, process, technology and market causes — without ranking them yet.

Use the template
Generate five distinct explanations for why BeautyCo's traffic
is increasing while conversion, satisfaction and repeat purchase
are declining. For each: state the cause, explain the logic,
identify supporting evidence, identify contradicting evidence,
identify additional data required.
Group into: Customer / Product / Process / People / Technology / Market.
Do not rank the explanations yet.
Evidence check

Use NotebookLM for qualitative evidence, Colab where numbers are relevant.

🧩 Pop quiz: which category does this belong to?
"Customers abandon checkout when a promo code fails to apply." Is this a Product, Process, or Technology cause?

Answer: Most teams call it a Technology bug — but it could just as easily be a Process gap (no fallback when a promo fails) or a Product issue (unclear promo terms). Good root-cause analysis holds more than one category open until evidence narrows it down.

Collaborate Board: "Post one plausible cause other groups may have missed."

🗳️ Nearpod Board
Suggested output (keep collapsed until the debrief)
Possible causes: Service Capacity Constraints (long queues, poor consultations) · Inventory Availability Issues (stock-outs, unavailable products) · Loyalty Experience Weakness (falling repeat purchases). Do not rank yet — Template 4 validates which are best supported.
Save your Root Cause Hypotheses.
Diagnose · Template 4

Critical Review & Evidence Validation

"Which explanation is best supported?"
~25 minNotebookLM + Colab + ChatGPT/Gemini
🧠 Critical Review & Evidence Validation (ReAct + Reflection)
Learn

The strongest point of integration in the course. Ground claims in documents, inspect the numbers, then run an AI critique.

Optional tool layer
Validation needTool / Agent
Internal evidence checkNotebookLM / Projects
External benchmarkDeep Research
Data pattern checkData Insight Agent
Structured critiqueImperial AI Council / Reflection pattern
1
Ground claims

What supports each hypothesis, what contradicts it, what is missing.

2
Inspect numerical evidence

Staff capacity vs. conversion, stock-outs vs. sales growth, store-level outliers.

3
Run an AI critique
Evaluate the hypotheses from Template 3 objectively. For each:
1. Supporting evidence 2. Contradictory evidence 3. Missing evidence
4. Possible bias/assumption 5. Confidence: High/Medium/Low
6. What evidence would change the conclusion.
Prioritise by evidence strength. Do not treat correlation as causation.

"How confident are you that staffing is a primary cause?" High / Medium / Low / Insufficient evidence.

🗳️ Vote
Suggested output (keep collapsed until the debrief)
HypothesisConfidence
Service Capacity ConstraintsHigh
Inventory Availability IssuesHigh
Loyalty Programme WeaknessMedium
Traffic Quality IssuesLow

Diagnosis statement: Traffic generation is not the primary issue. Conversion capability is.

Save your Prioritised Diagnosis, then continue to Decide.
Decide · Template 5

Options Assessment

"What could we do?"
~25 minChatGPT/Gemini + Google Sheets
🧠 Trade-off Evaluation (Multi-Perspective Reasoning)
Learn

Generate broadly, then evaluate rigorously — in three rounds: generate options, score against criteria, then examine from multiple stakeholder perspectives.

1
Generate options
Generate six possible actions to address the prioritised
BeautyCo diagnosis: low-cost quick wins, operational fixes,
customer-experience improvements, process/technology
improvements, longer-term strategic options. Do not recommend yet.
2
Evaluate options

Score: Impact · Feasibility · Cost · Risk · Evidence · Time-to-value.

3
Multiple perspectives
Evaluate each option from: Customer, Employees, Operations,
Finance, Brand, Risk & compliance. State what they value, would
support, concerns, trade-offs. Recommend a prioritised sequence.

"Which option should BeautyCo prioritise?" — vote before and after the evaluation.

🗳️ Run the Vote
Suggested output (keep collapsed until the debrief)
OptionImpactFeasibility
Increase Marketing SpendLowHigh
Improve StaffingMediumHigh
Improve Inventory PlanningMediumMedium
Integrated Service + Inventory ImprovementHighMedium

Recommended option: Integrated Service + Inventory Improvement — fix conversion capability before driving more traffic.

Save your Recommended Option, then continue to Deliver.
Deliver · Template 6

Executive Action Plan

"What will we do next?"
~20 minChatGPT/Gemini + Google Sheets
🧠 Action Planning (Plan-and-Solve / Prompt Chaining)
Learn

Translate the recommendation into coordinated execution using every previous output as context.

Framing Brief→ Situation Assessment→ Root Cause→ Diagnosis→ Recommended Option→ Action Plan
Use the template
Use all five previous outputs as context. Create a 30/60/90-day
executive action plan: recommended initiative, workstreams and
actions, owners, timeline, KPIs and targets, risks and mitigation,
dependencies, human approval points, stakeholder communication,
monitoring cadence. Identify what can be AI-assisted vs. requires
human judgement or approval.

Final reflection: strongest part of the plan, biggest unresolved risk, assumption to monitor.

🗳️ Final Reflection
Suggested output (keep collapsed until the debrief)

Key initiatives: Service Capacity — workforce scheduling, queue management, advisor training. Inventory Availability — inventory forecasting, automated replenishment, stock visibility. Loyalty Enhancement — personalized promotions, AI-assisted recommendations, tiered rewards.

KPICurrentTarget
Conversion18%22%
CSAT7180
Repeat purchase29%35%
Stock-out rate14%<8%
Save your Executive Action Plan — you've completed the full Decision Journey.
Wrap-Up

Key Takeaways & What's Next

Key Takeaways from the Decision Journey

  • Discover → Diagnose → Decide → Deliver maps directly onto six templates
  • NotebookLM grounds, Colab quantifies, ChatGPT/Gemini reasons — each at the right moment
  • Every template ends with a human challenge before saving the output
  • Each output becomes context for the next template

Coming Up

  • Review your full workbook at the BeautyCo Case Hub
  • Compare your recommendation with the suggested conclusion
  • Explore every tool used in the course at the AI Playground
Embedded AI thinking patterns
TemplateThinking Pattern
1. Decision FramingExpert Perspective (RGCTO)
2. Situation AssessmentStructured Analysis (Chain-of-Thought)
3. Root Cause AnalysisMultiple Explanations (Tree-of-Thought)
4. Evidence ValidationCritical Review & Evidence Validation (ReAct + Reflection)
5. Options AssessmentTrade-off Evaluation (Multi-Perspective Reasoning)
6. Executive Action PlanAction Planning (Plan-and-Solve / Prompt Chaining)

AI thinking patterns are embedded naturally — the focus stays on Decision Intelligence, not prompt engineering.

Ready to present your workbook?
0/6
Review the Full BeautyCo Case Hub →
Slide 1 of 8