Module 04V · Lesson 03
Decision Support and Strategic Synthesis
Reading time: 18 minutes Track: Role Path — Executive Leadership Prerequisites: Module 04V · Lessons 01 and 02
What this lesson does
Executives make decisions across many functions with varying depth of expertise. The work isn't usually "make the right call" — it's "make the call with enough context, soon enough, with appropriate inputs, after asking the right questions."
AI helps with this work substantially when used well. It can synthesize across documents you don't have time to read in detail, surface considerations you might miss, stress-test your initial position, and frame trade-offs you're navigating.
By the end of this lesson, you'll be able to:
- Use AI as a thinking partner for important decisions
- Apply specific frameworks for AI-assisted strategic synthesis
- Stress-test your own thinking before commitments
- Recognize when AI is helping vs. when it's just validating what you already believed
01 · The "thinking partner" stance
A specific reframe: AI as thinking partner means treating it as a sophisticated junior colleague — capable, available, but requiring direction and review.
Not: "Tell me what to do." Yes: "Help me think this through."
Not: "Make this decision." Yes: "What's the strongest case for and against this option?"
Not: "Generate the answer." Yes: "What questions should I be asking that I'm not?"
The framing matters because it preserves your role as decision-maker while extracting AI's value as analyst and synthesizer.
02 · Synthesis of long documents
You'll never read every document that crosses your desk. AI helps you process the volume.
The structured synthesis prompt
CEO synthesizing [document type] for [decision context]. The document is [length, source, purpose].
Provide:
1. Executive summary (3-5 sentences max)
2. Key facts I need to know
3. The 2-3 most important implications
4. Things in this document I should pay particular attention to
5. Questions this raises that I should ask my team
6. What's notably absent that should be addressed
[Document attached]
You read the synthesis. For consequential decisions, you also read the document. AI's synthesis is for triage, not replacement of your own engagement with critical material.
When AI synthesis is enough
- Background materials for general orientation
- Industry reports you want gist of
- Most board pre-reads (with selective deep reading)
- Routine financial reports
- News articles and intelligence
When AI synthesis isn't enough
- Material you'll quote or rely on in your own communications
- Highly technical material outside your expertise
- Material that's the basis for a consequential decision
- Material that might contain MNPI you need to be aware of
The discipline: know which category each document is in. Default to reading anything in the second category, even when AI offers a clean summary.
03 · Decision frameworks
For complex decisions, AI helps generate or stress-test frameworks.
The framework generation prompt
CEO facing a decision. The decision: [specific decision].
Context:
- [Background]
- [Constraints]
- [Stakes]
- [Time pressure]
Generate a decision framework that:
1. Identifies the key variables
2. Identifies the trade-offs between options
3. Identifies the information I'd need to decide well
4. Identifies the assumptions I'm making
5. Identifies my likely cognitive biases on this decision
6. Suggests how to structure the decision (single judgment, decision matrix, scenario analysis)
Be challenging. Don't validate my framing if there's a better one.
This produces a structured framework. Use as scaffolding for your own analysis.
The trade-off articulation prompt
For decisions with multiple competing considerations:
CEO weighing trade-offs between [Option A] and [Option B].
Option A: [description, benefits, costs]
Option B: [description, benefits, costs]
Help me think through:
1. What are the second-order effects of each option?
2. What does success look like for each? Failure?
3. Which option has higher variance? Higher expected value?
4. What's reversible? What's irreversible?
5. Under what conditions would I clearly prefer one over the other?
6. What information would change my preference?
7. What's the timing pressure on each option?
The output is a structured trade-off analysis. Verify against your own assessment; AI may miss organizational dynamics you know.
04 · The stress-test workflow
The highest-value executive AI use: stress-testing your own thinking.
The challenger prompt
CEO with a tentative decision. The decision: [your tentative position].
Reasoning: [why you think this is right].
Be a challenger. Specifically:
1. What's the strongest case against this position?
2. What am I likely underweighting?
3. What information would change my mind?
4. What assumptions am I making that might not hold?
5. What's the worst plausible outcome of going forward with this decision?
6. What would a sophisticated critic outside my organization say?
7. What's the "we should have seen this coming" failure mode?
Don't soften the critique. Take the contrarian position seriously.
The output is uncomfortable. Engage with it seriously.
What to do with the output
Triage:
- Real concerns you hadn't considered — address them before deciding
- Concerns you'd considered and dismissed — revisit; the dismissal may have been right or wrong
- Reaches that aren't actually substantive — note and dismiss honestly
- Concerns where you accept the risk — document explicitly that you considered and accepted
The discipline isn't accepting every challenge. It's engaging with every challenge seriously.
The sycophancy trap
AI defaults to agreeing with you (from Module 01). The challenger prompt overrides this. But you need to actually engage; you can't override the prompt back to validation.
Watch for: AI surfacing a real concern, you saying "but here's why that's not actually a problem," AI agreeing. That cycle is sycophancy. The right pattern is AI surfacing a concern, you engaging substantively, the concern getting either addressed or honestly dismissed.
05 · Cross-functional translation
You navigate constantly between functions. Each has its own vocabulary, concerns, and decision criteria. AI helps with the translation work.
Specific translations
Science → Finance:
"Translate this Phase 2 result into financial implications. Specifically: probability of Phase 3 success based on this readout, likely Phase 3 cost and timeline, NPV implications, runway impact."
Regulatory → Commercial:
"Translate this FDA guidance into commercial implications. Specifically: what does this mean for our launch plan, label expectations, pricing/access strategy?"
Operations → Strategy:
"Translate this operations status into strategic implications. Specifically: does this change our timeline for [milestone], does it change our resource needs, does it change our partnership strategy?"
Each translation surfaces issues that wouldn't be obvious to the originating function. That's the value of executive cross-functional thinking; AI accelerates it.
06 · Scenario planning
A specific use case where AI excels: generating and exploring scenarios.
The scenario generation prompt
CEO planning for [decision / strategic context]. Help me develop scenarios.
Current situation: [description]
Generate:
1. Base case scenario (most likely outcome)
2. Upside scenario (favorable conditions)
3. Downside scenario (unfavorable conditions)
4. Black swan scenario (low probability, high impact)
For each scenario:
- Triggering conditions
- Likely outcomes
- Strategic response options
- Specific actions we'd take
- What we'd want to have done in advance
Don't manufacture scenarios for the sake of having four. Be specific about what would drive each.
This produces structured scenario thinking. Use as input to strategic planning conversations.
When scenario planning is most valuable
- Strategic financing decisions
- Major partnership opportunities
- Phase 3 design decisions
- Pre-launch planning
- M&A targets and timing
- Pipeline prioritization
Less valuable for routine operational decisions where scenarios collapse to "go" or "don't go."
07 · The "questions to ask" framework
A specific high-value pattern: using AI to generate the questions you should be asking.
Why this matters
Executives often face decisions in functions where they don't have deep expertise. The risk isn't making the wrong decision — it's not asking the right questions to inform the decision.
CEO preparing for [meeting / decision] in [function area]. My expertise here is limited.
Generate the 10-15 questions I should be asking my team to make sure I'm informed before making this decision. For each question, briefly note:
- Why it matters
- What a strong answer looks like
- What a weak answer would reveal
This produces a structured question list. You use it in your team meeting. Your questions are sharper than they would be without prep.
A variant: the "what's missing" framework
After your team has briefed you, before making a decision:
CEO finishing briefings on [decision]. The team has covered:
- [Topic 1]
- [Topic 2]
- [Topic 3]
What's likely missing from this briefing? What questions should I ask before deciding?
This catches gaps. Particularly useful for areas outside your expertise.
08 · A worked example
A realistic scenario: a CEO considering a strategic financing.
Setting: Your runway is 14 months. Your data readout is in 18 months. Standard advice is to finance with 24 months of runway. Three options: (1) traditional follow-on; (2) PIPE with strategic investor; (3) non-dilutive partnership.
Step 1 — Decision framework
Use the decision framework prompt. Get back structured framework covering capital efficiency, signaling, control, optionality, timing.
Step 2 — Trade-off articulation
Run the trade-off prompt for each pair (follow-on vs. PIPE, PIPE vs. partnership, etc.). Structure the trade-offs explicitly.
Step 3 — Stress-test
Use the challenger prompt on your initial preference (let's say: PIPE with strategic investor). AI surfaces:
- Signaling risk (PIPE may be read negatively)
- Concentration risk (strategic investor influence)
- Optionality cost (foreclosing partnership)
- Alternative: blended approach
Engage with each seriously. Two of the four are real concerns you'd not fully considered.
Step 4 — Questions for advisors
Generate the questions to ask your CFO, your board, your IR advisor, your lawyer. Take the list into each conversation.
Step 5 — Decision
After the conversations, you decide. The decision is yours. AI helped structure your thinking.
Total time: ~3-4 hours of AI-assisted thinking spread over a week of conversations. Without AI: similar total time, less structured thinking, less stress-tested decision.
09 · Knowledge check
Three questions.
Q1. Why is the "thinking partner" framing recommended over "tell me what to do"?
a) AI can't tell you what to do b) Treating AI as thinking partner preserves your role as decision-maker while extracting AI's value as analyst and synthesizer — "tell me what to do" abdicates the executive judgment that's irreducibly yours c) Thinking partners are cheaper d) The framing doesn't matter
Q2. What's the most important discipline when using AI to stress-test your thinking?
a) Always accept AI's critiques b) Engage with each challenge seriously — triage to "real concerns you'd missed," "concerns you'd considered and dismissed," "reaches not substantive," "accepted risks documented" — but don't override the prompt back to validation c) Only stress-test small decisions d) Avoid stress-testing because it slows decisions
Q3. Why is the "questions to ask" framework called out as high-value?
a) Executives are bad at asking questions b) Executives often face decisions outside their expertise; the risk isn't making the wrong decision but not asking the right questions — AI-generated question lists make your team briefings significantly more useful c) It's a way to avoid making decisions d) It substitutes for AI itself
Answers: Q1: b · Q2: b · Q3: b
10 · What's next
Lesson 04: Building AI fluency in your organization — your role as a leader, not just a user.
End of Lesson 03.