Learn with Phoebe Β· Applied AI

Learn AI PPT with Phoebe

Six sessions, 45 minutes each, and one real deck you build across all of them. AI can make a presentation in thirty seconds. It cannot tell you whether the room understood it. This course teaches the part that decides: the 30-second test, the one slide method with peer-reviewed evidence behind it, and an honest read on which presentation advice you have been handed is research and which is just repeated a lot.

6sessions
1real deck you keep
27% understood at the start
91% by session 6
45min per session
A darkened boardroom. A beautifully designed chart glows on the projection screen while five of the six people in the front row look down into their phones, each face lit from below. One person at the end is still looking up.
The deck was never the problem. Five faces lit from below, one still looking up. Every tool in this course can produce the slide on that screen. None of them can produce the sixth person.

The six sessions 🎬 · one real deck, built across all of them

Pick a deck you genuinely have to present in the next month. Not a practice deck. Every session moves that one forward, and the order is deliberate, because the mistakes compound backwards: people design slides before they have an argument, and then no amount of design saves it.

⏱️ Session 1 · easy

The 30-second test

Thirty seconds after your slide goes up, can the room say what you are asking for? Meet the test, watch a real slide climb from 27 percent understood to 91, and learn the method that has actual research behind it.

β–Ά Start here
🧱 Session 2 · easy

Outline first

Your deck is an argument. Generate the structure, then kill half of it. Why every good AI deck tool puts an editable outline gate in front of you, and why the bad ones skip straight to slides.

the argument
🎯 Session 3 · medium

Slides that carry

Assertion-evidence in full: a sentence headline that states the message, one visual that proves it. One idea per slide, one hero number, and the three Mayer principles that decide whether any of it lands.

biggest lever
πŸ“Š Session 4 Β· medium

Charts and data

Charts that argue instead of decorate, and where fabricated figures actually come from. Microsoft says Copilot is not optimised for verbatim reproduction of large inputs. That sentence is why your number drifted.

integrity
πŸ› οΈ Session 5 Β· hands-on

Build the deck

The real build, tool by tool, with the licence and export traps named: Copilot needs a Designer licence for half its features, Gamma flattens gradients, Canva warns your PPTX may not survive PowerPoint. Plus a course still selling a tool that no longer exists.

the real thing
πŸ—£οΈ Session 6 Β· hands-on

The room

Speaker notes, the Q and A you are dreading, and the accessibility pass no AI tool will do for you. This is where the last nine points of the score live, and no vendor claims WCAG conformance for its output.

πŸ”’ unlocks the ceiling
easy - everyone medium hands-on hardest

Your passport 🎫

Clear all three quiz questions in a session and it gets stamped. Four stamps unlocks the accessibility lever in the simulator, which is the only way to reach 91 percent. There is an honest bypass if you would rather skip ahead.

Choose your path πŸ—ΊοΈ

Three doors into the same conviction: a slide earns a decision, or it is a document that wasted a room's afternoon.

πŸ”₯ Presenting this week 1β†’ 3β†’ 6 the test, the headline rewrite, then the accessibility and Q and A pass
🧭 Building a deck from nothing 2β†’ 3β†’ 5 argument first, then slides, then whichever tool you have a licence for
πŸ“ The whole journey 1β†’ 2β†’ 3β†’ 4β†’ 5β†’ 6 27 to 91, and the ceiling is only reachable at the end

What is honest here 🧾

Three things this course will not do, said up front.

ClaimStatus
Assertion-evidence beats topic-plus-bullets on comprehensionPeer-reviewed. Garner and Alley 2013, 110 students, significant at p under .01
The 10/20/30 rule is optimalConvention. Kawasaki's own reasoning, asserted from experience, no study behind it
The simulator's comprehension percentageA teaching model of the research above, not a measurement of your room. The contrast ratio in it is real WCAG maths

The knowledge map 🧠

The whole course at a glance - hover a session to spotlight its concepts, click any node to jump in.