Compression makes you efficient. Amplification makes you strategic. These four operations extend your thinking capacity — generating more options, finding more weaknesses, reaching more audiences, and seeing more consequences than you could alone.
The Four Operations
ACTIVE
◇Diverge
Generate multiple options
MANUALYou think of 2-3 approaches based on your experience
AI-AMPAI generates 5-8 approaches drawing on patterns you haven't considered, then you select and combine the best elements
◈Pressure-Test
Find weaknesses before stakeholders do
MANUALYou anticipate objections based on past experience, but miss blind spots
AI-AMPAI systematically attacks your strategy from every stakeholder's perspective in minutes
◆Translate
Adapt one strategy for multiple audiences
MANUALYou rewrite the same deck 3 times for different stakeholders
AI-AMPAI generates audience-specific framings from a single source of truth — same strategy, different emphasis per audience
◉Project
Map second-order consequences
MANUALYou think one or two steps ahead but rarely have time to map full consequence chains
AI-AMPAI traces ripple effects across dimensions: viewer behavior, advertiser adoption, platform trust, creator incentives, competitive position
⚡
The Amplification Principle
Your judgment stays central. AI handles the generative labor — producing options, finding weaknesses, adapting language, tracing consequences. You handle the evaluative labor — deciding which options are right, which weaknesses matter, which framing lands, which consequences to prioritize.
Ready-to-Use Prompt Templates
Each prompt is designed for AI video quality work but generalizes to any strategic design project. Copy, fill in the brackets, and use with the Stop Chatting, Start Directing framework from Level 1.
Your AI video quality framework involves multiple stakeholder perspectives. Pressure-testing lets you pre-empt challenges instead of discovering them in review meetings.
The Stakeholder Gauntlet
Systematically attack your strategy from each stakeholder's position
Here is my current framework for [specific aspect of AI video quality]: [paste framework]
Act as three different stakeholders, one at a time, and critique this framework from their perspective:
1. ENGINEERING LEAD: What's technically infeasible, underspecified, or will create scaling problems? Where does this framework assume capabilities we don't have?
2. ADS PRODUCT MANAGER: Where does this conflict with advertiser needs, revenue goals, or adoption friction? What would make advertisers reject or ignore these standards?
3. POLICY/LEGAL: What creates regulatory exposure? Where could this framework fail under GDPR, FTC advertising guidelines, or platform transparency requirements?
For each perspective, give me: the top 3 objections ranked by severity, and one specific revision that would address the strongest objection.
Format as three distinct sections I can use to pre-brief before my next review.
The Inversion Test
Find what breaks by imagining the opposite
Here is my proposed quality standard for [specific element — e.g., text hooks in AI video ads]: [paste standard]
Invert this: Describe what an AI video ad looks like that technically PASSES this standard but is still a bad ad. What loopholes exist? What does "malicious compliance" look like?
Then: propose 2-3 specific tightening revisions that close the most dangerous loopholes without making the standard so rigid it blocks good creative work.
The Pre-Mortem
Assume the framework failed and work backwards
Here is my quality framework for AI-composed video ads: [paste or summarize]
It's 6 months from now and this framework has failed. Adoption is low, quality hasn't improved, and stakeholders are frustrated.
Give me 5 distinct failure scenarios — not vague risks, but specific causal chains. For each:
- What went wrong first
- Why it cascaded
- What early warning sign we should have seen
- One preventive action we could take now
Rank them by likelihood, not severity.
You're operating across engineering, product, legal, and creative teams. Each group needs to understand the same quality strategy — but they care about different things and speak different languages. One source of truth can generate multiple translations.
The Audience Adapter
Generate stakeholder-specific versions of the same strategy
Here is my design strategy for [AI video quality / specific initiative]: [paste core strategy document]
Generate three versions of a one-page summary of this strategy, each tailored to a different audience:
VERSION 1 — ENGINEERING: Lead with technical architecture and implementation requirements. Emphasize what needs to be built, system constraints, and quality signals that can be measured programmatically. Use precise technical language.
VERSION 2 — BUSINESS/ADS PRODUCT: Lead with advertiser impact and revenue implications. Emphasize how this improves ad performance, reduces advertiser churn, and creates competitive advantage. Use business metrics and outcomes language.
VERSION 3 — EXECUTIVE (VP+): Lead with strategic positioning and competitive landscape. Emphasize why this matters now, what the risk of inaction is, and how this connects to company-level priorities. Keep it to 5-7 bullet points max, each one a decision-ready statement.
Each version should be standalone — someone reading only their version should fully understand the strategy and their role in it.
The Objection Translator
Convert one team's concern into another team's language
I received this feedback from [engineering/legal/product]: "[paste the specific feedback or objection]"
I need to communicate this concern to [different team]. Translate it into language and framing that will resonate with their priorities:
- What's the equivalent concern in their world?
- What evidence or framing would make them take it seriously?
- What's the ask — what specifically do I need from them?
Keep it to 3-4 sentences I can use in a Slack message or meeting.
Your AI video quality work sits at an intersection where decisions ripple outward — what you define as 'quality' shapes advertiser behavior, viewer experience, creator incentives, and platform trust simultaneously.
The Consequence Map
Trace ripple effects of a strategic decision across dimensions
I'm considering this strategic decision for AI video ads: [describe the decision — e.g., "requiring all AI-generated video ads to pass a minimum coherence score before serving"]
Map the second and third-order consequences across these five dimensions:
1. VIEWER BEHAVIOR: How does this change what viewers see, how they engage, and what they expect over time?
2. ADVERTISER RESPONSE: How do advertisers adapt? Do they invest more in quality, find workarounds, or leave the platform?
3. CREATOR/TOOL ECOSYSTEM: How does this affect the AI tools advertisers use to create content? What does the tool market do in response?
4. PLATFORM TRUST: How does this affect Meta's position on AI content quality relative to competitors?
5. INTERNAL PRECEDENT: What standard does this set for other AI-generated content beyond ads?
For each dimension, give me: the immediate effect (0-3 months), the adaptation effect (3-12 months), and the equilibrium state (12+ months).
Flag any dimension where the short-term and long-term effects conflict.
The Fork in the Road
Compare two strategic paths with full consequence analysis
I'm deciding between two approaches for [specific decision]:
OPTION A: [describe]
OPTION B: [describe]
For each option, analyze:
- Who wins and who loses (across viewer, advertiser, engineering, and platform perspectives)
- What it makes easy and what it makes hard going forward
- The irreversible commitments it creates
- The optionality it preserves
Then: is there an Option C that captures the best elements of both? If so, describe it. If not, explain why the tradeoff is genuine and which option you'd recommend given [state your priorities].
You're already doing option generation, but you can push it further. Instead of asking for 'different approaches,' you can constrain the divergence to produce genuinely distinct options — not just variations on the same idea.
The Constraint Shift
Generate options by changing the underlying assumption, not just the surface approach
I'm working on [specific design challenge — e.g., "how to evaluate hook effectiveness in AI video ads"].
Generate 4 fundamentally different approaches by shifting the core constraint each time:
1. OPTIMIZE FOR SPEED: What if we need this solved in 2 weeks with minimal engineering investment?
2. OPTIMIZE FOR RIGOR: What if accuracy matters more than speed and we have 2 months?
3. OPTIMIZE FOR SCALE: What if this needs to work across 50 markets and 12 languages from day one?
4. OPTIMIZE FOR LEARNING: What if the primary goal isn't a solution but generating data that helps us make a better decision in 6 months?
For each, give me: the approach in 2-3 sentences, what it sacrifices, and what it uniquely reveals that the others don't.
The Borrowed Lens
Apply frameworks from adjacent domains to your problem
Here's my current design challenge: [describe the AI video quality problem you're working on]
Apply three different strategic lenses from outside the ads/design domain:
1. EDITORIAL LENS: How would a newspaper editor or film studio approach quality standards for this content?
2. INFRASTRUCTURE LENS: How would a civil engineer or urban planner think about setting standards that need to scale and evolve?
3. MARKETPLACE LENS: How would an economist or market designer think about incentive structures that produce quality as an emergent property?
For each lens: what's the key insight, what specific principle would they apply, and how does it translate into a concrete recommendation for my context?
Putting It Together — Your Weekly Amplification Rhythm
MONRun a Consequence Map on your biggest open decision. Takes 15 min. Gives you a strategic view for the week.
WEDBefore any stakeholder review, run the Stakeholder Gauntlet on whatever you're presenting. Pre-empt 80% of pushback.
THUUse the Audience Adapter to generate stakeholder-specific versions of anything you're sharing cross-functionally.
FRIRun a Borrowed Lens or Constraint Shift on one problem you're stuck on. Fresh perspective with zero meetings.
~1 hour/week of amplification work → strategic leverage that used to take days