Training Public Speaking with AI

Like any behavioral skill, public speaking only improves with repetition and feedback. But rehearsing in front of real people costs time and availability, and we don't always dare. A conversational AI fills this gap: it helps sharpen the content, anticipate difficult questions, and receive structured feedback, as many times as you want. This is the "AI + psychology" angle at the heart of the Kognytia approach.

Three ways to use AI to improve

graph LR
    A["1. Speech architect<br/>(structure & clarity)"] --> D["More persuasive<br/>speaking"]
    B["2. Simulated audience<br/>(questions & objections)"] --> D
    C["3. Feedback coach<br/>(script/transcript analysis)"] --> D

1. The architect: the AI helps clarify the single message, build the narrative arc and tighten the content.
2. The simulated audience: the AI plays an audience that asks the hard questions, to prepare the dreaded Q&A.
3. The coach: the AI analyzes your script or the transcript of your rehearsal and gives precise feedback.

Prompt 1 — Clarify the message and structure

"I have to give a [duration] presentation to [describe the audience] on [topic]. My goal is for them to [intended action]. Help me: (1) phrase a single message in one sentence of fewer than 15 words, (2) propose a punchy hook, (3) organize the content into three pillars following a problem → solution narrative arc, and (4) write a strong closing sentence with a call to action. Ask me questions if you're missing information."

Prompt 2 — The audience that asks the hard questions

The Q&A is often the most stressful moment. Train it:

"Here is the summary of my presentation (paste it). Play the role of a skeptical and demanding audience: ask me, one at a time, the 10 hardest questions or objections I might face. Wait for my answer to each question, then briefly assess how solid my answer was and suggest a better phrasing if needed."

This training removes the element of surprise: on the day, you'll have already met most of the questions.

Prompt 3 — Feedback on the rehearsal

Rehearse out loud, transcribe your delivery (or dictate it), then ask for an analysis:

"Here is the transcript of a rehearsal of my speech (paste the text). Analyze: (1) the clarity of the main message, (2) the presence and quality of the hook and conclusion, (3) the passages that are too long or confusing to cut, (4) the filler words and verbal tics, (5) the ethos/pathos/logos balance. Finish with a score out of 10 and the three priority improvements."

Table: what AI does well… and its limits

Keep in mind: AI prepares the substance and the anticipation, but the delivery (voice, body, presence) is built standing, ideally filmed or in front of a few people. Combine the two.

Precautions

Don't feed confidential information (customer data, sensitive internal figures) into an AI assistant without caution; anonymize. And keep your own voice: AI helps structure and correct, but a speech that sounds "too polished" and impersonal convinces less than authentic speech. Use its suggestions as raw material, not as a finished product.

Practical exercise

For your next talk, chain the three prompts: first have the structure clarified, then train the Q&A on 10 questions, finally submit the transcript of a rehearsal for feedback. Note the three priority improvements, apply them, and do a final rehearsal standing and out loud (without AI) to work on delivery. Goal: arrive on the day on familiar ground.

Summary

AI is a powerful training partner for public speaking: architect of the message and structure, simulated audience to prepare the dreaded Q&A, and coach for feedback on your rehearsals. Three ready-to-use prompts cover these uses. But AI mainly works the substance: voice, body and presence are rehearsed standing and out loud. Anonymize your data and keep your authentic voice.