What's a "good presentation" now that AI can generate polished slides and a script from a one-line brief?

The shift

Turning an idea into a polished deck — clean layout, on-brand visuals, a coherent narrative arc, and a word-for-word script — collapses from scarce design-and-writing hours to a one-line brief and a few seconds. What stays scarce is having something worth presenting, owning the argument well enough to defend it live, and the presenter's genuine command of the material in the room.

The axioms

• A polished, well-designed deck signals the presenter took the work seriously and understands it — production effort is proof of care and competence.
• Building a good deck is slow and expensive, so the deck is the deliverable people spend their prep time on — production is the scarce bottleneck.
• A tight narrative arc — the right sequence, the setup, the reveal — is a scarce craft skill that separates strong communicators from weak ones.
• The person who can make the argument look good in slides is trusted to have the strongest argument — presentation quality proxies for thinking quality.
• Standing up to present is how a group decides whose idea to back — the meeting is where alignment and commitment happen.
• A presenter fielding hard questions live demonstrates they actually understand what they put on the slides — Q&A is the verification step.
• Good speakers are made by rehearsal, feedback, and reps — delivery skill is scarce and slow to build.
• The audience's attention and synchronous time are worth gathering people for — a live presentation is worth the room's collective hour.

Invalid axioms

1. A polished, well-designed deck signals the presenter understood and cared about the work. Layout, visual polish, and a clean narrative structure were expensive enough that producing them was evidence of investment and grasp. A model now generates all of it from a sentence. Habit-trap: audiences and reviewers still read production quality as a proxy for rigor, and presenters still spend their scarce prep hours polishing slides instead of stress-testing the argument.
2. Building a good deck is the slow, expensive part of preparing. Design, copy, and sequencing were the bottleneck that ate the prep time. That cost is now near-zero. Habit-trap: teams still schedule days for "deck production" and treat a finished deck as the milestone, when the finished deck no longer proves the thinking behind it exists.
3. A tight narrative arc is a scarce craft skill. Setup, tension, reveal, and clean transitions are exactly the pattern-matched structure a model produces well. Habit-trap: still prizing and hiring for "storytelling" as a standalone production skill, when the arc is now generatable and the scarce part is whether there's a true, defensible claim inside it.
4. The person who makes the argument look best in slides has the strongest argument. When looking good was costly, polish correlated with substance because only the committed paid for it. That correlation breaks when polish is free. Habit-trap: rooms still award the decision to the best-produced deck, so a thin idea in a great template now beats a strong idea in a plain one.

Unchanged axioms

1. Having something worth saying stays scarce. A model can produce a deck about anything; it can't supply the insight, the decision, or the reason the room should care. Deciding what is worth an audience's attention is a taste-and-judgment problem, not a production one, and it got no cheaper.
2. The presenter must own the argument well enough to defend it live. Q&A, pushback, and the follow-up question that isn't on any slide are real-time verification. A presenter who commissioned slides they don't understand fails here in a way the deck can't cover for — and being confidently wrong under questioning is more exposing now that the slides themselves prove nothing.
3. Command of the room stays human. Reading the audience, sensing the objection nobody voiced, adjusting the argument on the fly, and holding a group's trust through disagreement is relationship and judgment work a generated script doesn't touch. The scarce thing was never the slides; it was the person in front of them.
4. Trust comes from an accountable presenter who knows the material. The claim on the slide needs someone who will answer for it, act on it, and be held to the commitment made in the room. That accountability doesn't get cheaper because the slide production did.
5. A live gathering is worth the room's time only when it does something a document can't. Deciding, negotiating, aligning, building conviction — these still justify pulling people together. What no longer justifies it is broadcasting information a generated summary could have delivered async.

New axioms

1. Polished decks now routinely mask thin thinking. When a one-line brief yields a professional deck, production quality and argument quality fully decouple, and audiences have lost the cue they used to calibrate trust. Distinguishing a good argument from a good-looking deck becomes the reviewer's actual job, and most rooms aren't set up to do it.
2. Presenters can't defend slides they didn't think through. AI-generated decks let someone present material they don't genuinely understand, and the gap only surfaces under live questioning — if anyone asks. Orgs need a way to locate the accountable understanding behind a deck, because the artifact no longer carries it.
3. The meeting and the deck become ritual when the content is free. If a deck can be generated on demand and read async, the standing presentation-and-review meeting risks persisting as theater — a slot people fill because the calendar says so, not because a decision needs a room. What earns synchronous time has to be re-justified.
4. Verification load moves from producing to interrogating. When anyone can generate a convincing deck in minutes, an org can produce more polished, authoritative-looking arguments than any audience has time to pressure-test. The scarce act shifts from making the case to auditing whether the case is real, and nobody clearly owns that.

Where it breaks

Rooms still award the decision to the best-produced deck (invalid) at the moment polish fully decouples from substance (new) — so the deck that wins is now selected on exactly the signal that stopped meaning anything, and thin ideas in strong templates beat strong ideas presented plainly. Separately, teams still treat the finished deck as the prep milestone (invalid) while the only remaining verification is a presenter defending the argument live (new) — prep time is going into polishing generatable slides instead of the one thing the deck can no longer prove, so the gap between a presenter and their material stays hidden until someone in the room asks the question that isn't on a slide.

ai-axioms.vercel.app/axioms/marketing/comms-presentations-ai-axioms