Claude Prompts for Talks That Actually Land
20 copy-paste Claude prompts for keynotes, panels, conference talks, and pitch presentations. Use Claude's long context to feed entire transcripts of past talks for consistency coaching.
In short: This page contains 17 copy-paste ready prompts, organized into 5 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Hand-curated and tested by the AI Academy team.
Talk Structure + Writing
4 promptsTalk Outline from Topic
1/17✨ What it does
Claude builds a talk outline on [describe] for [describe] at [X minutes], with a 60-second hook, 3-5 sections each anchored by a story, transitions, a closing callback, and a final ask. Skip the agenda-slide opener, then lock the stories you can actually tell.
Build a talk outline. Topic: [describe]. Audience: [describe]. Length: [X minutes]. Format: keynote / panel / breakout. Output: hook (first 60 seconds — what stops scrolling brains), 3-5 sections with 1 anchoring story each, transitions between sections, callback at end, final ask. Avoid the "agenda slide" opener.
Pro tip: Per-section: 1 story + 1 takeaway > 3 abstract points. Audiences forget bullets, remember stories. Each section needs a memory anchor.
Hook Opener (First 60 Seconds)
2/17✨ What it does
Claude writes 5 hook openers under 90 words for a talk on [topic] for [describe]: a contrarian claim, a vivid scene, a question, a surprising stat, and a personal failure, tagged by audience. Say each hook out loud, then open with the one that does not need extra setup from you.
Write 5 hook openers for a talk on [topic]. Audience: [describe]. Each hook: under 90 words, attention-grabbing, lands without context. Variations: contrarian claim, vivid scene, audience question, surprising stat, personal failure. Tag which works for which audience.
Pro tip: Skip "thank you for having me" + "let me start by." Audiences decide in 60 seconds whether you're worth listening to. Earn attention before earning trust.
Story Mining from Transcripts
3/17✨ What it does
Claude mines [Paste 2-3 full talk transcripts I've given] for 8-12 stories you have already told, each with a one-line premise, payoff, the point it makes, and when you last used it, in a table. Paste real transcripts, then pick unused stories for the next talk.
[Paste 2-3 full talk transcripts I've given]. Mine my actual stories I've told before. Extract: 8-12 stories I've used, each with — premise (1 line), payoff (1 line), what point it makes, when last used. Output as table I can pick from for new talks.
Pro tip: Claude's 200K context shines here. Feed it 10 past talks; get a story bank. Speakers reuse stories more than they realize — knowing your inventory prevents accidental repeats AND under-use of strong material.
Closer + Call to Action
4/17✨ What it does
Claude writes 3 closers under 60 seconds for a talk ending on [main message], each calling back the opener, stating one action, and landing emotionally without a last-word thank you. Rehearse all three, then end on the message, not the thanks.
Write 3 closer variations for a talk ending on [main message]. Each closer: under 60 seconds, references the opening hook (callback), states one specific action, leaves audience emotionally landed (not flat). Avoid "thank you" as the final word — let the message land last.
Pro tip: Last sentence = most remembered. "Thank you" wastes that real estate. End on the line you want quoted. Then nod and walk off; applause fills the silence.
XML tags are just the start. Learn the full Claude workflow.
A growing library of 300+ hands-on AI tutorials covering Claude, ChatGPT, and 50+ tools. New tutorials added every week.
Audience + Context
4 promptsAudience Research Brief
5/17✨ What it does
Claude writes a research brief for [conference / event / audience] covering who attends, why they show up, what they have already heard, what prior speakers said, expected language, and hot buttons. Read it before you write slides, then cut anything that restates common knowledge.
Research brief for [conference / event / audience]. Output: who attends (demographics, roles), why they attend (jobs to be done), what they've heard before (avoid restating common-knowledge), what speakers before me said (don't contradict accidentally), language register expected, hot buttons + sacred cows.
Pro tip: Find last year's talks (often on YouTube). Knowing what the audience already heard = avoid restating their priors. They came to learn something new, not hear the same keynote remixed.
Adapt 1 Talk to 3 Audiences
6/17✨ What it does
Claude adapts [Paste talk outline] for [audience 1], [audience 2], and [audience 3], saying which sections to cut, which stories to swap, which jargon to change, and how the hook shifts. Keep the core message, then rebuild the rest for the room you are actually in.
[Paste talk outline]. Adapt this talk for: (A) [audience 1 — describe], (B) [audience 2], (C) [audience 3]. Per audience: which sections cut, which stories swap, which jargon adjust, what hook changes. Keep core message; change everything else around it.
Pro tip: Same core message + different stories = the secret to giving 20 talks a year without burning out. Don't rewrite; remix.
Cultural Context Check
7/17✨ What it does
Claude flags cultural context for [talk on X] in [country/region]: examples that will not translate, jokes that may offend, business norms versus the US, religious or political sensitivities, and prep timing. Cut the risky jokes, then replace your examples with ones that work there.
I'm giving [talk on X] in [country/region]. Flag cultural context I should know: examples that won't translate, jokes that won't land or might offend, business norms different from US, religious or political sensitivities, time zone for prep timing. Specific to this region — not generic "be respectful."
Pro tip: US speaker examples (sports analogies, US-centric stats, "Founding Fathers" callbacks) frequently flop abroad. Swap for local references. Audiences notice when speakers haven't bothered to localize.
Translator/Interpreter Brief
8/17✨ What it does
Claude briefs a simultaneous interpreter for your talk on [topic], with 30-50 terms, idioms that need substitutes, name and acronym pronunciation, pause points, and jokes that need a warning. Send the brief before the event, then pause where it marked.
Brief a simultaneous interpreter for my talk on [topic]. Output: 30-50 specialized terms with intended meaning, idioms I use that don't translate (with neutral substitutes), names + acronyms pronunciation guide, places where I'll pause for translation, jokes that need pre-warning.
Pro tip: Send the brief 1 week ahead. Surprise interpreters = bad translation. Fed-context interpreters = your talk lands in 5 languages.
Q&A + Panels
4 promptsQ&A Prep — Hardest Questions
9/17✨ What it does
Claude generates the 15 hardest, most adversarial questions for [talk on X], each with why they would ask, 3 response strategies, what not to say, and a one-sentence ideal answer. Rehearse the one-sentence answers, then avoid the lines it told you not to say.
I'm giving [talk on X]. Generate the 15 hardest, most uncomfortable, most adversarial questions an audience could ask. Per question: why someone would ask it (what they're really after), 3 response strategies, what NOT to say, the 1-sentence ideal answer.
Pro tip: Prep for hostile, not friendly. Friendly questions handle themselves. Knowing your weakest 15 questions in advance = composure when one lands live.
Panel Talking Points
10/17✨ What it does
Claude preps panel talking points for [describe] with other panelists [list with bios] and your role [describe]: 5 unique points of view, 3 stories, 2 contrarian takes, and one question to redirect. Keep the points only you can offer, then drop a story if the table goes bland.
Panel prep. Topic: [describe]. Other panelists: [list with bios]. My role: [describe]. Output: 5 unique POVs only I can offer (don't overlap with others), 3 stories ready to drop, 2 contrarian takes if conversation gets bland, 1 question I want to redirect to. Keep me distinctive.
Pro tip: Panels bore audiences when everyone agrees. Have 1-2 contrarian takes ready. Don't pick fights — pick angles others won't cover. Distinct = memorable.
Bridge Lines for Going Off-Topic
11/17✨ What it does
Claude writes 8 smooth bridge lines under 15 words that redirect questions back to [my main message] without dismissing the asker, as a cheat sheet you can rehearse. Practice the lines until they sound casual, then use one the first time a question drifts.
Write 8 bridge lines for redirecting questions back to [my main message]. Each line: smooth (not dismissive), under 15 words, lets me pivot without ignoring the question. Examples: "That's exactly the question — and the deeper issue is..." Output as a cheat sheet I can rehearse.
Pro tip: Politicians have 5 stock bridge lines memorized. Speakers should too. The point isn't evasion — it's steering toward your strongest material when off-topic questions hit.
Hostile Question Recovery
12/17✨ What it does
Claude walks through handling a hostile on-stage question like [describe], with a 3-second composure ritual, a validate-without-agreeing move, a reframe, a brief answer, a redirect, and body-language notes. Rehearse the composure beat, then pivot to your strongest ground.
Walk through how to handle a hostile question on stage. Question type: [describe — e.g., "your industry is destroying X"]. Output: 3-second composure ritual, validation move (acknowledge without agreeing), reframe to your strongest ground, brief response, redirect to next question. Body language notes.
Pro tip: Acknowledge before disagreeing. "I hear that — and here's why I see it differently" reads as confident. "You're wrong because..." reads as defensive. Same content, different audience reaction.
These prompts give you the what. Tutorials give you the why.
Learn when to use extended thinking, how to build Claude Projects, and workflows that compound. 300+ tutorials and growing.
Delivery + Practice
4 promptsRehearsal Plan
13/17✨ What it does
Claude builds a 7-day rehearsal plan for a [length-min] talk in [venue type], with daily tasks, times, and what not to change in the last 48 hours. Follow the day list, then freeze your talk when it says to stop rewriting.
Build a 7-day rehearsal plan for a [length-min] talk in [venue type]. Day 1-7 specific tasks: outline lock, full read-throughs, recorded run, feedback session, room rehearsal, light prep, day-of warmup. Time per day. What NOT to change in last 48 hours.
Pro tip: Lock content 48h before. Last-minute rewrites read as flustered, not polished. Use final 48h for delivery only — the words are already chosen.
Recorded Run Critique
14/17✨ What it does
Claude critiques [Paste transcript of my recorded rehearsal] for filler words, sentences too long to breathe, energy drops, jokes that fail on the page, where to slow down, and where to cut, with timestamps if you tagged them. Paste a real run, then cut the slowest section first.
[Paste transcript of my recorded rehearsal]. Critique honestly: filler words count, sentences too long for breath, places I lost energy, jokes that didn't land in writing (revisit delivery), where to slow down, where to cut. Specific timestamps if I tag them.
Pro tip: Recording reveals what you can't hear live. "Um" count, sentence drag, energy dips. Brutal feedback from a transcript is cheaper than brutal feedback from an audience.
Energy + Pacing Map
15/17✨ What it does
Claude builds an energy and pacing map from outline [paste], scoring each section 1-10, slow, medium, or fast, pause points, still versus moving, and where to drop volume. Mark those levels in your script, then vary energy so the talk does not go flat.
Build an energy + pacing map for my talk. Outline: [paste]. Output: per section, target energy level (1-10), pacing (slow/medium/fast), pause moments, mic stance (still/moving), where to drop volume for emphasis. Map prevents flat-line delivery — energy varies, audience stays awake.
Pro tip: Flat energy = lost audience by minute 8. Variation = the thing TED speakers train on most. Quiet moments make loud moments louder.
Day-Of Warmup Routine
16/17✨ What it does
Claude writes a time-blocked day-of warmup for [time of talk], covering night-before sleep and screens, morning, 5-minute voice and body warmups, visualization, food, and the last 30 and 5 minutes. Run your last 30 minutes as written, then stop tinkering.
Write a day-of warmup routine for [time of talk]. Output: night-before sleep + screen rules, morning routine, voice warmup (5 min), body warmup (5 min), mental rehearsal (visualization), what to eat (and avoid), 30 minutes before stage, 5 minutes before stage. Specific, time-blocked.
Pro tip: Voice warmup matters more than people think. Cold cords = cracking, breathy, low projection. 5 minutes of humming + tongue twisters = audible difference.
Long-Context Coaching
1 promptVerbal Tic Pattern Audit
17/17✨ What it does
Claude audits verbal tics across [Paste 3-5 of my full talk transcripts], ranking fillers, hedging like I think or kind of, repeated openers and closers, and recycled sentences. Highlight the top filler, then catch it in your next rehearsal.
[Paste 3-5 of my full talk transcripts]. Audit my verbal tics across all of them. Output: filler words ranked by frequency, hedging phrases I overuse ("I think", "kind of"), repeated openers/closers, sentences I keep recycling. Show me what an audience hears across multiple talks.
Pro tip: You can't hear your own tics in real time. 5 transcripts in Claude = patterns visible. Most speakers have 3-5 unconscious crutches; killing them = instantly more polished.
Free tool
Prompt Optimizer
Turn a rough idea into a structured, professional AI prompt.
Frequently Asked Questions
Prompts are the starting line. Tutorials are the finish.
A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.
7-day free trial. Cancel anytime.
Related guides