There are two ways to work with your deck's AI designer. You can let it make the creative decisions, or you can direct them yourself. Neither is wrong — but most people over-direct, and it actually makes their decks worse.
This article explains why trusting the AI first and steering second produces better results than trying to specify everything upfront.
The AI Already Has Good Taste
When you say "add a traction slide," the AI doesn't just throw your numbers on a blank page. It reads your brand, your content, and the type of deck you're building, then makes dozens of design decisions — layout, emphasis, spacing, motion, colour intensity — based on what it knows about effective slides.
A single revenue number gets dramatic treatment: large, centred, with space to breathe. Four team members get a grid layout with enough room for each person. A list of features gets structured cards with visual hierarchy. The AI makes these choices because it understands what works — not randomly, but based on your content.
You don't need to direct those decisions. You can just react to the result.
Three Modes of Working
Think of these not as quality levels, but as modes of collaboration. Each one is right at different moments.
Mode 1: Let the AI Lead
"Add a team slide."
The AI handles everything — how many people to feature, the layout, the visual style, the interactions. It makes opinionated choices based on your brand and the type of deck you're building.
Use this when: You're drafting, exploring, or you genuinely want to see what the AI comes up with. This is the fastest way to build, and the results are often better than you'd expect — because the AI is making design decisions you might not have thought of.
Mode 2: React and Steer
"This feels too corporate — warm it up."
"Make the revenue number the first thing you see."
"I love what you did with the team slide — apply that energy to the closing slide too."
You look at what the AI built and respond with your reaction. You're not specifying layout or animation — you're describing the impression you want, and the AI translates that into design choices.
Use this when: The AI's first pass is close but not quite right. This is where most great slides get made — through two or three rounds of steering, not through one perfect specification.
Mode 3: Direct a Specific Choice
"Put the metric in the top left with a comparison chart beside it."
"Use a 2x2 grid with circular photos."
You tell the AI exactly what to do on a particular element. This overrides the AI's judgment on that specific choice while letting it handle everything else.
Use this when: You have a clear picture in your head of how something should look, and the AI's interpretation isn't matching it. This is an override, not the default way of working.
Why Over-Directing Hurts Your Deck
Here's the counterintuitive thing: specifying every detail often produces worse results than trusting the AI and steering.
You constrain the AI's best ideas. When you say "2x2 grid, circular photos, name underneath, hover reveals bio," the AI executes exactly that. But it might have created something better — maybe an asymmetric layout that makes your CTO more prominent because they're the key hire, or a design that fits your brand's personality better than a standard grid.
You fight the AI instead of collaborating with it. Long, detailed specifications often conflict with the AI's design instincts. The AI knows your brand colours work better with certain spacing and contrast ratios. When you override all of that, the pieces don't fit together as naturally.
You spend time on decisions the AI handles well. Choosing between a grid and a timeline, or deciding on animation timing, or picking the right hover effect — these are things the AI does well based on your content. Spend your time on what only you know: which metric matters most, what impression you want to make, what your audience cares about.
The Sweet Spot
The best decks come from this rhythm:
- Let the AI lead — get the slide on the page
- React to what you see — "this feels too busy" or "the headline doesn't grab me"
- Steer the impression — "make it feel more confident" or "the investor should feel urgency"
- Direct only when needed — "actually, put the timeline horizontal across the top"
Most slides are done after step 2 or 3. You rarely need step 4, and when you do, it's for one specific element — not the whole slide.
What This Means in Practice
A slide that doesn't feel right
Instead of rewriting a detailed specification, just describe what's off:
"This is too generic — it could be anyone's slide. Make it feel like ours."
"There's too much on this slide — simplify it to just the key point."
"The numbers are there but they don't feel impressive. Make the growth rate hit harder."
The AI knows how to respond to these because they describe an outcome, not a method.
A slide you love
Tell the AI what works. This is just as valuable as telling it what doesn't:
"This is exactly right — the spacing, the tone, everything. Keep this energy for the rest of the deck."
"The way the metrics are presented here is perfect. Use this approach for the financial slides too."
Positive steering teaches the AI what resonates with you, which improves every slide that follows.
A slide with a specific vision
When you have a very clear picture in your head, it's fine to direct:
"I want the closing slide to be just our logo, centred, on a dark background. Nothing else. Let it breathe."
This is a creative choice only you can make. The AI executes it. But notice you're still describing an impression ("let it breathe") alongside the specific direction — you're collaborating, not just instructing.
What to Read Next
- Iterative Refinement: The 3-Prompt Rule — A practical workflow for steering a slide from good to exceptional in three rounds.
- Steering Your Deck's Design Direction — How to describe the impression you want without over-specifying.
- Prompting for Tone and Personality — The vocabulary of feeling — words that shape how your slides land.
There are two ways to work with your deck's AI designer. You can let it make the creative decisions, or you can direct them yourself. Neither is wrong — but most people over-direct, and it actually makes their decks worse.
This article explains why trusting the AI first and steering second produces better results than trying to specify everything upfront.
The AI Already Has Good Taste
When you say "add a traction slide," the AI doesn't just throw your numbers on a blank page. It reads your brand, your content, and the type of deck you're building, then makes dozens of design decisions — layout, emphasis, spacing, motion, colour intensity — based on what it knows about effective slides.
A single revenue number gets dramatic treatment: large, centred, with space to breathe. Four team members get a grid layout with enough room for each person. A list of features gets structured cards with visual hierarchy. The AI makes these choices because it understands what works — not randomly, but based on your content.
You don't need to direct those decisions. You can just react to the result.
Three Modes of Working
Think of these not as quality levels, but as modes of collaboration. Each one is right at different moments.
Mode 1: Let the AI Lead
"Add a team slide."
The AI handles everything — how many people to feature, the layout, the visual style, the interactions. It makes opinionated choices based on your brand and the type of deck you're building.
Use this when: You're drafting, exploring, or you genuinely want to see what the AI comes up with. This is the fastest way to build, and the results are often better than you'd expect — because the AI is making design decisions you might not have thought of.
Mode 2: React and Steer
"This feels too corporate — warm it up."
"Make the revenue number the first thing you see."
"I love what you did with the team slide — apply that energy to the closing slide too."
You look at what the AI built and respond with your reaction. You're not specifying layout or animation — you're describing the impression you want, and the AI translates that into design choices.
Use this when: The AI's first pass is close but not quite right. This is where most great slides get made — through two or three rounds of steering, not through one perfect specification.
Mode 3: Direct a Specific Choice
"Put the metric in the top left with a comparison chart beside it."
"Use a 2x2 grid with circular photos."
You tell the AI exactly what to do on a particular element. This overrides the AI's judgment on that specific choice while letting it handle everything else.
Use this when: You have a clear picture in your head of how something should look, and the AI's interpretation isn't matching it. This is an override, not the default way of working.
Why Over-Directing Hurts Your Deck
Here's the counterintuitive thing: specifying every detail often produces worse results than trusting the AI and steering.
You constrain the AI's best ideas. When you say "2x2 grid, circular photos, name underneath, hover reveals bio," the AI executes exactly that. But it might have created something better — maybe an asymmetric layout that makes your CTO more prominent because they're the key hire, or a design that fits your brand's personality better than a standard grid.
You fight the AI instead of collaborating with it. Long, detailed specifications often conflict with the AI's design instincts. The AI knows your brand colours work better with certain spacing and contrast ratios. When you override all of that, the pieces don't fit together as naturally.
You spend time on decisions the AI handles well. Choosing between a grid and a timeline, or deciding on animation timing, or picking the right hover effect — these are things the AI does well based on your content. Spend your time on what only you know: which metric matters most, what impression you want to make, what your audience cares about.
The Sweet Spot
The best decks come from this rhythm:
- Let the AI lead — get the slide on the page
- React to what you see — "this feels too busy" or "the headline doesn't grab me"
- Steer the impression — "make it feel more confident" or "the investor should feel urgency"
- Direct only when needed — "actually, put the timeline horizontal across the top"
Most slides are done after step 2 or 3. You rarely need step 4, and when you do, it's for one specific element — not the whole slide.
What This Means in Practice
A slide that doesn't feel right
Instead of rewriting a detailed specification, just describe what's off:
"This is too generic — it could be anyone's slide. Make it feel like ours."
"There's too much on this slide — simplify it to just the key point."
"The numbers are there but they don't feel impressive. Make the growth rate hit harder."
The AI knows how to respond to these because they describe an outcome, not a method.
A slide you love
Tell the AI what works. This is just as valuable as telling it what doesn't:
"This is exactly right — the spacing, the tone, everything. Keep this energy for the rest of the deck."
"The way the metrics are presented here is perfect. Use this approach for the financial slides too."
Positive steering teaches the AI what resonates with you, which improves every slide that follows.
A slide with a specific vision
When you have a very clear picture in your head, it's fine to direct:
"I want the closing slide to be just our logo, centred, on a dark background. Nothing else. Let it breathe."
This is a creative choice only you can make. The AI executes it. But notice you're still describing an impression ("let it breathe") alongside the specific direction — you're collaborating, not just instructing.
What to Read Next
- Iterative Refinement: The 3-Prompt Rule — A practical workflow for steering a slide from good to exceptional in three rounds.
- Steering Your Deck's Design Direction — How to describe the impression you want without over-specifying.
- Prompting for Tone and Personality — The vocabulary of feeling — words that shape how your slides land.