AI Prompts for ChatGPT for Public Speaking: Structure, Storytelling, Delivery, Anxiety Management

I am preparing [TALK: CONFERENCE KEYNOTE / WEDDING SPEECH / ALL-HANDS / MEDIA INTERVIEW / INVESTOR PITCH / BOARD PRESENTATION] for an audience of [AUDIENCE: SIZE, BACKGROUND, WHY THEY ARE IN THE ROOM]. Help me find the one sentence I want each person to be able to repeat to someone else twenty-four hours later. Ask me: what did I initially plan to talk about, what does this audience already know that I do not need to repeat, what does this audience specifically not want to hear (the cliches, the assumptions, the corporate speak), and what am I actually here to say. Reflect back the one message in one sentence with the words the audience will remember.

My one message is [ONE MESSAGE]. Audience is [AUDIENCE]. Talk length is [LENGTH: 5 MINUTES / 15 MINUTES / 30 MINUTES / 45 MINUTES]. Design the structure. Not a five-point list. Recommend the specific structure that fits my message and audience: the problem-agitation-resolution arc, the three-act story spine, the counterintuitive-thesis-with-evidence structure, the personal-story-into-broader-point structure, or the specific hybrid that this material actually calls for. Justify the choice, then map my message onto that structure with the section labels and the transition logic.

The opening ninety seconds of my [TALK TYPE] have to earn me the right to keep talking. My audience is [AUDIENCE], my one message is [ONE MESSAGE], and I want to open with [OPENING TYPE: A STORY / A COUNTERINTUITIVE CLAIM / A QUESTION / A SPECIFIC OBSERVATION ABOUT THIS ROOM]. Write three versions of the opening ninety seconds in my voice. Each must land me in my body physically, capture attention without being cheesy, and set up the through-line of the talk. Do not use the phrase good morning everyone or thank you for having me.

The three core points in my talk are [POINT ONE, POINT TWO, POINT THREE]. Each one needs a story that makes it land, because audiences remember stories and forget bullet points. For each point, generate three story options drawn from my own life or work that I can retrieve. Ask me the surfacing questions if I have not thought of them. Then help me pick the strongest one per point and structure it into the seventy-five second story-tell format that lands in a talk: setup, complication, moment, meaning.

The closing of my talk has to leave the audience with my one message and the specific action or shift I want them to carry. My message is [ONE MESSAGE], the action I want is [ACTION: A SPECIFIC THING THEY DO DIFFERENTLY / A SPECIFIC REFRAME THEY HOLD / A SPECIFIC ASK]. Write three versions of the closing ninety seconds. No cliches, no thank-you-for-having-me, no rushed disclaimers. The closing must land the emotional register and give them the specific handle to walk out with.

Here is the outline and full talk for [EVENT]: [PASTE OUTLINE OR FULL TEXT]. Rehearse it with me by walking through the talk section by section and telling me: what this specific audience will be thinking at each section, what part might lose them, where the energy will peak and dip, where I am likely to rush from adrenaline, and where the pause before the next section is worth an extra beat. This is directorial feedback, not a rewrite.

After my talk on [TOPIC] to [AUDIENCE], there will be a Q&A. Simulate it with me. Generate the fifteen most likely questions from this specific audience, ordered by likelihood, including: the friendly softball, the genuine curiosity, the underneath-question they will not phrase directly, the pushback from the skeptic, the derail (someone using Q&A to make their own point), and the question I most dread. For each, give me the three-sentence response frame in my voice.

The hardest moment in my talk is transitioning from [SECTION A] to [SECTION B]. It is where I lose the room in rehearsal. Help me nail it. Options: a story that bridges the two sections, a specific claim that sets up the next section, a callback to the opening, a pause and reset. Give me three options for the specific bridge language, and tell me which is likely to land best for this audience.

My talk is [LENGTH]. Under adrenaline I tend to [ADRENALINE PATTERN: SPEED UP AND FINISH FIVE MINUTES EARLY / LOSE THE PLOT IN THE MIDDLE / CUT THE CLOSING TO SAVE TIME / EXPAND THE OPENING AND CRAM THE ENDING]. Help me plan the timing markers I install into the talk so pressure does not distort it. Include: the section-by-section time gate (I should be at this section at this minute mark), the specific pace cue I give myself if I am running fast, the recovery move if I am running long, and the practice discipline in rehearsal that installs the actual pace.

For [EVENT] on [DATE], design the run-of-show prep sequence over the seven days leading up to the talk. Day-by-day plan including: final content lock, out-loud rehearsals (spaced not crammed), the specific rehearsal audience (one friend, one colleague from the audience type, a video recording), the pre-day mental prep, and the day-of pre-talk sequence from waking up through walking on. Realistic, not obsessive. The goal is trained and calm, not overprepared and rigid.

When I am preparing for or delivering [TALK TYPE], my anxiety shows up as [PATTERN: CANNOT SLEEP THE NIGHT BEFORE / RACING THOUGHTS IN THE HOUR BEFORE / DRY MOUTH AND SHAKING HANDS ON STAGE / MIND BLANKS MID-TALK / RUSH AND FINISH TOO FAST / AVOID THE EYES OF THE AUDIENCE]. Diagnose the specific pattern. Physiological (adrenaline expression), cognitive (specific story running), or behavioral (learned avoidance). For each specific expression, name the mechanism and the specific intervention that addresses that expression, not generic breathwork advice.

For the two hours before [TALK TYPE], design my pre-talk physical protocol. Include: what to eat (or not) two hours before, hydration (specific amount, timing), the movement that discharges nervous energy (specific and doable in whatever green room I have), the breath pattern that regulates from that state to calm alert (specific inhale-exhale count), the voice warm-up that does not embarrass me if a stagehand walks in, and the physical anchor I hold in the last thirty seconds before walking on.

The ten seconds before I walk on stage or into the meeting or on the video call is where the anxiety spikes hardest. Design the specific ten-seconds-before protocol. Include: the physical posture I take, the breath pattern (one specific breath), the internal phrase I say (three to five words maximum), the eye contact or gaze I hold, and the first physical action I take when the moment begins. Rehearsable, so that in the actual moment I execute the protocol rather than trying to decide what to do.

The specific fear I have about [TALK] is that my mind will blank mid-talk. Design the mid-blank recovery protocol. Include: the physical reset (pause, breath, ground contact), the internal reframe (a blank is a pause, the audience will forgive it if I do not panic), the specific move (return to the last thing I remember saying, restate the section title, ask the audience a rhetorical question that buys me thirty seconds), and the mental frame that a small blank is normal at every level of speaking including experienced keynote speakers.

After [TALK], design my post-talk debrief so I learn from the actual performance rather than either rehashing endlessly or dismissing the experience. Within twenty-four hours: the specific what actually happened versus what I feared happened, the three things I want to remember for the next time, the one thing I want to change, and the honest thank-you note to myself for showing up. Not a performance review, a learning ritual.

I have a full [FULL LENGTH] version of my talk on [TOPIC]. I now need to give it in [SHORTER LENGTH: 5 MINUTES / 15 MINUTES / 10 MINUTES]. Help me adapt. Not compress: cut. Identify: the one message that must survive the cut, the one story that carries it, the parts to remove entirely (not shorten), and the specific new transitions that stitch the shorter version so it lands as a complete talk rather than as a rushed version.

I gave [TALK] to [ORIGINAL AUDIENCE]. Now I am giving a version of it to [NEW AUDIENCE: DIFFERENT INDUSTRY / DIFFERENT SENIORITY / DIFFERENT CONTEXT]. Adapt it. Identify: what parts of the original talk assumed knowledge this new audience does not share, what parts of the original talk this new audience already knows so I can compress them, the new opening that lands with this specific audience, the story swap where the original story does not translate, and the new closing calibrated to what this audience needs to walk out with.

I have given [TALK ON TOPIC] as a keynote. Now I am doing a media interview on the same topic (podcast / TV segment / print interview). Adapt the material. Include: the ninety-second version of my thesis for a first question, the two-sentence version for a shorter format, the three anecdotes I have in my back pocket for different question directions, the specific pivot I use when a question misses the point, and the specific line I want the interviewer to quote if I can only leave one.

I have given [TALK] and want to turn it into a written piece: an essay, an article for [PUBLICATION], a LinkedIn post series, or a chapter. Turn it. Include: the honest structural change from spoken to written (spoken repetition removed, denser argument, cleaner transitions), the additional evidence a written reader will want that a live audience did not, the specific hook line that works as an article opening rather than as a talk opening, and the format-specific tightening for the piece length.

Turn this session into the reusable public speaking prep workflow I run for every talk from now on. Include: the one-message discovery protocol, the structure decision tree, the opening and closing generation, the story surfacing questions, the Q&A simulation, the pre-talk physical and mental protocols, the ten-second walk-on protocol, and the post-talk debrief. Single document, three pages maximum, so I have the process on paper for every future talk without reinventing it.