A checklist is only useful if you can actually run it, item by item, and finish more confident than you started. This one is built for that. It is organized by the stages of producing a generated deck, from before you press generate to the moment before you present, and every item carries a one-line reason so you know why it earns a place rather than treating it as ritual.
Use it as a working gate, not a reading exercise. Keep it open while you build a deck and tick items off as you clear them. The items are deliberately concrete, so each one is either done or not done, with no ambiguity about whether you have satisfied it.
Skip items on purpose when the stakes are genuinely low, but never skip them by accident. The point of a checklist is to make the omission a decision rather than an oversight.
One note on how to use the stages. Earlier items prevent later problems, which is why the list is ordered the way it is. A weak outline caught before generation saves an hour of rework; a fabricated number caught before presenting saves your credibility. The cost of a missed item grows the further down the list it sits, so resist the temptation to rush the early, cheap-feeling items in order to get to the part that feels like real work.
Before You Generate
Define the Outcome
- [ ] You can finish the sentence "after this deck, the audience will ___" with a verb. Without a target action, the deck has no spine and the tool optimizes for nothing.
- [ ] You have named the audience and what they already know. Context is what separates a sharp deck from generic filler.
Build the Structure
- [ ] You wrote the section outline by hand before prompting. Structure is the part the tool cannot do for you, and it determines draft quality.
- [ ] The section order makes your case, not just a logical survey. A deck that informs without persuading misses the goal.
- [ ] You decided roughly how many slides the deck needs. An unbounded prompt produces sprawl, and trimming a bloated draft costs more than scoping it up front.
- [ ] You know which single slide carries your most important point. If you cannot name it, the deck has no center of gravity and the audience will not find one either.
During Generation
Frame the Prompt
- [ ] Your prompt leads with goal and audience before the outline. Context up front produces dramatically better output, as shown in Turning a Rough Outline Into Finished Slides, Step by Step.
- [ ] You generated once and read it before regenerating. Rerolling without reading chases randomness instead of fixing problems.
The Truth Pass
Verify Every Claim
- [ ] Every number, date, and name is checked against a real source. Generated facts are confident regardless of accuracy.
- [ ] No placeholder or invented statistic survives. One exposed fabrication can sink your credibility, the core warning in Where Generated Decks Go Sideways, and What Fixes Them.
- [ ] Any commitment, price, or timeline reflects what you actually agreed to. Anywhere a deck touches money, accuracy is the whole job.
The Editing Pass
Cut and Tighten
- [ ] Each slide makes one point in a few words. Dense slides compete with your narration and lose the audience.
- [ ] You deleted more than you added. The value in a generated deck is editorial subtraction.
- [ ] Transitions between slides hand off cleanly. The tool optimizes slides, not the journey between them.
- [ ] No slide is a transcript of what you plan to say. If the slide contains your whole sentence, the audience reads ahead and stops listening.
- [ ] The deck's voice sounds like you, not like generic generated prose. A read-cold rewrite of any flat line is the cheapest way to restore your tone.
The Brand Pass
Make It Yours
- [ ] Your colors, fonts, and logo are applied throughout. The default theme signals generic effort.
- [ ] Generic stock visuals are replaced with specific ones or removed. Concrete imagery reads as care.
Before You Present
Final Gate
- [ ] You read the full deck aloud once. Reading aloud surfaces unsayable lines and broken transitions silent review misses.
- [ ] Every slide serves the outcome from the first item. Cut any that do not. The disciplined version of this gate lives in Habits That Separate Polished AI Decks From Sloppy Ones.
- [ ] You tested the exported file in the environment where you will present it. Fonts and formatting can shift between the editor and the projector, and you do not want to discover that live.
- [ ] You can state the deck's single takeaway in one sentence. If the deck has done its job, that sentence is what the audience walks away repeating. The reusable model behind this whole sequence is laid out in The Draft-Shape-Refine Model for Generated Slides.
Adapting the Checklist Over Time
Retire Items That Never Catch Anything
A checklist should evolve with your skill. If you have internalized a habit so thoroughly that an item never catches a problem anymore, retire it deliberately to keep the list lean. A bloated checklist gets skimmed; a tight one gets used. The goal is a list short enough that you actually run every item.
Add Items From Your Own Failures
Every time a deck goes out with a flaw the checklist did not catch, add an item that would have caught it. Your checklist should encode the specific mistakes you tend to make, not just generic ones. Over time it becomes a personalized record of the lessons you have already paid for once.
Keep One Version Per Deck Type
A proposal, a conference talk, and an internal update stress different stages. Rather than one generic list, keep lightly tailored versions that weight the relevant passes, the truth pass for proposals, the editing and brand passes for talks. The shared structure stays; the emphasis shifts to match the deck in front of you.
Frequently Asked Questions
How long should running this checklist take?
For a short deck, ten to fifteen minutes beyond your normal editing, most of it in the truth pass. The before-you-generate items take only a couple of minutes but save the most time downstream, because they prevent the structural rework that eats hours.
Which items should I never skip, even when rushed?
The truth pass items. Verifying numbers and removing fabrications is the one stage where skipping carries outsized risk, since a single bad fact can undo the entire deck in front of an audience. Everything else degrades gracefully; this one fails catastrophically.
Can I use this for an internal, low-stakes deck?
Yes, but trim it consciously. For an internal update you might skip the brand pass and the read-aloud gate while keeping the truth pass. The rule is to drop items as a decision, never by accident, so you always know what you traded away.
Why are the before-you-generate items first?
Because they have the highest leverage. The quality of a generated deck is largely decided before you press generate. Getting the outcome, audience, and structure right up front means the rest of the checklist clears quickly, while skipping them makes every later item harder.
Should the checklist change for different deck types?
The stages stay the same; the emphasis shifts. A proposal weights the truth pass heavily because of pricing and commitments. A conference talk weights the editing and brand passes. Keep the full list and adjust how much scrutiny each stage gets to the deck at hand.
How do I keep this from becoming mindless ritual?
Read the one-line reason on any item you are tempted to rubber-stamp. The justifications exist so each tick is a real judgment, not a reflex. If an item never catches anything for your work, retire it deliberately rather than ticking it on autopilot.
Key Takeaways
- Run this as a working gate, ticking items off while you build, not as a passive reading.
- The before-you-generate items, outcome, audience, structure, carry the highest leverage on final quality.
- The truth pass is the one stage to never skip; a single fabricated fact can sink the whole deck.
- Editing is mostly subtraction; aim to delete more than you add and keep one point per slide.
- Trim items deliberately for low-stakes decks, but never skip them by accident.