The Visual Metaphor Generator: Turn Abstract Concepts Into Pitch-Deck-Ready Analogies
Why this prompt matters
A board or investor who doesn't grasp the mechanism behind a technical improvement won't advocate for the budget behind it, and a pitch that gets rejected or misunderstood because of a weak analogy costs far more than the 15 minutes it takes to generate better ones — teams that lead with the wrong metaphor often end up re-explaining the same concept across three follow-up meetings instead of moving the conversation forward.
What we use it for
You have 4 minutes on the agenda at next week's board meeting to explain why the company's new fraud-detection model reduces false declines, and the board has finance and legal backgrounds with zero machine learning experience — a slide full of precision and recall numbers will lose them in the first 20 seconds.
Prompt
You are a presentation coach and communications strategist who has helped hundreds of founders and executives explain complex ideas to boards, investors, and cross-functional teams. CONTEXT: I need to explain a concept in an upcoming presentation, and a plain technical or business explanation isn't landing with my audience. Concept to explain: [YOUR CONCEPT — e.g., "how our recommendation algorithm reduces churn"]. Audience: [WHO YOU'RE PRESENTING TO — e.g., "a board of directors with finance backgrounds, no technical depth"]. Context of the presentation: [BOARD MEETING / INVESTOR PITCH / ALL-HANDS / SALES DEMO / OTHER]. Time I have to explain this: [30 SECONDS / 2 MINUTES / 5+ MINUTES]. TASK: 1. Generate 5 distinct visual metaphors that explain [YOUR CONCEPT] to [AUDIENCE]. 2. For each metaphor, write: the one-sentence version I'd say out loud, a 2-3 sentence expanded version for when I have more time, and what visual I could put on a slide to reinforce it, described precisely enough that a designer could build it. 3. For each metaphor, rate its risk of falling flat or being misunderstood by this specific audience, and explain why. 4. Recommend which single metaphor to lead with and why it's the best fit for this audience and time constraint. CONSTRAINTS: - Do not reuse the same source domain (sports, cooking, driving, etc.) for more than one metaphor — I want 5 genuinely different angles, not 5 variations on one theme. - Avoid metaphors that require the audience to already understand a different complex concept to get the metaphor. - Flag any metaphor that could be seen as reductive or condescending to this specific audience. - Keep the one-sentence version under 25 words — it needs to work spoken aloud, not read off a slide. OUTPUT FORMAT: For each of the 5 metaphors: ## Metaphor [N]: [Source domain, e.g., "Air traffic control"] **One-sentence version:** [...] **Expanded version:** [...] **Slide visual:** [...] **Risk of falling flat:** [LOW/MEDIUM/HIGH] — [why] End with: **Recommended lead metaphor:** [which one] — [1-2 sentence justification tied to the specific audience and time constraint]
Result
## Metaphor 1: Airport security screening
One-sentence version: Our fraud model is like adding a second, faster security lane that still catches the same threats but stops fewer honest travelers for extra screening.
Expanded version: Right now, every transaction goes through one screening process that's cautious to the point of flagging too many legitimate customers. This model adds a smarter first pass — most legitimate transactions go straight through, while suspicious ones still get the full check. The threats caught stay the same; the false alarms drop.
Slide visual: A simple two-lane airport diagram — one lane labeled "current process" with a long queue and several false stops, one lane labeled "new model" showing a shorter queue with only genuinely flagged bags pulled aside.
Risk of falling flat: LOW — airport security is a universally understood, low-controversy reference point for this board.
## Metaphor 2: Credit scoring precision
One-sentence version: We tightened the model's judgment the same way a lender sharpens a credit score to approve more good borrowers without approving more bad ones.
Expanded version: A blunt credit score rejects some good borrowers just to be safe. A sharper model separates good and bad borrowers more precisely, so the lender can approve more good applicants at the same risk tolerance. Our fraud model does the same for transactions.
Slide visual: A simple two-axis chart — risk tolerance held constant on one axis, approval rate for legitimate customers increasing on the other, with a labeled improvement arrow.
Risk of falling flat: LOW — this board has finance backgrounds, so credit scoring is an unusually strong fit, arguably the strongest of the five.
## Metaphor 3: Weather forecasting confidence
One-sentence version: Think of it like a weather forecast that got more confident — same data, but now it can tell a 60% chance of rain from a 95% chance instead of calling both "maybe rain."
Expanded version: A vague forecast forces you to prepare for rain every time there's any chance. A precise one lets you skip the umbrella most days and only prepare when it really matters. Our model went from vague fraud "maybe" signals to sharper, more confident ones.
Slide visual: Two weather forecast icons side by side — a fuzzy "chance of rain" cloud icon versus a precise percentage readout.
Risk of falling flat: MEDIUM — clear concept, but weather forecasting doesn't map as tightly to a financial decision as the credit scoring metaphor does for this specific audience.
Recommended lead metaphor: Credit scoring precision — it's the only one of the five that speaks directly in the vocabulary this board already uses daily (approval rates, risk tolerance, false positives and negatives), so it requires zero translation step before the point lands, which matters given the 4-minute time constraint.
Most "explain this simply" prompts produce one metaphor and stop there, which means you're betting the whole presentation on whatever the model generated first. This prompt is built around a different assumption: the first metaphor that comes to mind is rarely the best one for a specific audience, and the only way to find the right one is to generate several genuinely different options and compare them side by side before you're standing in front of the board.
Why the prompt forces five different source domains
The constraint against reusing the same source domain — no two metaphors both drawn from sports, for instance — exists because left unconstrained, a model tends to generate small variations on its first idea rather than genuinely different angles. Five metaphors that are all sports analogies aren't five options, they're one option restated. Forcing five distinct domains (in the example output: airport security, credit scoring, and weather forecasting) surfaces real alternatives, and different audiences respond to different domains for reasons that have nothing to do with how clear the metaphor is in the abstract — a board of finance people will find a credit-scoring metaphor more immediately legible than an equally clear sports analogy, simply because it's already in their working vocabulary.
The risk rating is the part most metaphor-generation prompts skip
A metaphor that's technically accurate can still fail in the room — because it's condescending, because it requires background knowledge the audience doesn't have, or because it accidentally implies something you didn't intend. Asking the model to rate each metaphor's risk of falling flat, and explain why, forces it to reason about audience fit rather than just generating plausible-sounding analogies and leaving the judgment call entirely to you. This is also why the prompt asks for the audience and presentation context up front — a metaphor that works in a casual all-hands can read as flippant in front of investors, and the model can only flag that mismatch if it knows the setting.
Why both a one-sentence and expanded version
Presentations rarely go exactly as rehearsed. The one-sentence version is what you say if you get interrupted after ten seconds or the meeting runs long and your slot gets cut to a third of its planned length; the expanded version is what you use if you have the full time and want to build the metaphor out with more texture. Having both prepared in advance means you're not improvising a compressed version live under time pressure, which is when explanations tend to get muddled.
The slide visual instruction matters for a specific reason
The prompt asks for a visual described precisely enough that a designer could build it — not just "a relevant image." A vague instruction like "show a chart representing improvement" produces a slide that looks like every other slide in the deck. Asking for a specific, buildable visual (a labeled two-lane diagram, a two-axis chart with a specific labeled arrow) gives you something concrete to hand to whoever builds your slides, or to sketch yourself in five minutes, rather than another abstraction layer you still have to solve.
Adapting it for your situation
The time-constraint field changes what a good recommendation looks like more than any other input. A metaphor that's excellent for a 5-minute deep-dive slot can be the wrong choice for a 30-second elevator answer, because it needs the expanded version to land. When you fill in the bracketed time field honestly — not optimistically — the model's final recommendation will actually account for the fact that you might only get to say one sentence before someone asks a follow-up question.