What You Can Do with AI
Real tasks. Real prompts. Copy, paste, and get results in minutes.
Turn any document into a video series — without being on camera
How to use this
- Run this prompt in ChatGPT (chatgpt.com) or Claude (claude.ai) to generate the video scripts from your document.
- Go to synthesia.io. Create a new video. Select an AI avatar and voice that matches your brand.
- Paste each VIDEO script into Synthesia's scene editor — one scene per section. Add the on-screen text suggestions where indicated.
- Generate and download all 3 videos. Upload to your LMS, YouTube, or internal wiki. You now have a professional training series without ever being on camera.
Turn this document into a 3-video training series optimized for AI avatar delivery (Synthesia). DOCUMENT DETAILS: - Title: [DOCUMENT TITLE — e.g. "New Employee Onboarding Policy"] - Target audience: [WHO WILL WATCH — e.g. "New hires in their first week"] - Key goal: [WHAT THEY SHOULD KNOW/DO AFTER WATCHING — e.g. "Understand our PTO policy, know how to submit requests, avoid common mistakes"] VIDEO 1 — Orientation (2 minutes, 3 scenes): What this is, why it matters, and what they'll learn. VIDEO 2 — Walkthrough (3 minutes, 4–5 scenes): Step-by-step how to do it. Each scene covers one action. VIDEO 3 — Troubleshooting (3 minutes, 5 scenes): Top 5 mistakes people make and how to fix each one. SCRIPT RULES (optimized for AI avatar delivery): - Short sentences only (max 15 words per sentence) - One idea per scene - End each scene with: ON-SCREEN TEXT: "[key takeaway in 5–8 words]" - Include [PAUSE — 1 second] markers between major points - No jargon, idioms, or complex vocabulary - Write for spoken delivery: contractions are fine, conversational tone - Include [VISUAL CUE: description] for any graphics or diagrams to show on screen DOCUMENT TO CONVERT: [PASTE YOUR FULL DOCUMENT HERE — policy, process doc, FAQ, training manual, product guide, etc.]
Why this works: One document becomes three professional training videos. ChatGPT writes the scripts optimized for spoken delivery, Synthesia handles the avatar, voice, and production. You never touch a camera. Onboarding, compliance, product demos — all from a document you already have.
Create brand visuals without a designer
Midjourney
Leonardo AI·Course 10 — AI Mood Boards & Style Guides + Course 48 — Fire Your Graphic DesignerHow to use this
- Use PROMPT 1 to build your brand's visual world. Fill in each of the 8 layers. Run it in Midjourney (midjourney.com > /imagine) or Leonardo AI (leonardo.ai > Image Generation).
- Once you have a visual style you love, use PROMPT 2 for specific assets — thumbnails, social posts, ads. Fill in the platform, emotion, and brand details.
- Generate 4 variations. Pick the best one, upscale it, and download.
- Use the outputs directly for social media, website headers, pitch decks, or mood boards. Save your 8-layer prompt as a template for future visual consistency.
--- PROMPT 1: The 8-Layer Architecture (run in Midjourney or Leonardo AI) --- Fill in each layer, then combine them into one prompt: LAYER 1 — Subject/Brand World: [WHAT YOU'RE VISUALIZING — e.g. "A premium sustainable skincare brand visual world"] LAYER 2 — Emotional Tone: [HOW IT SHOULD FEEL — e.g. "calm and confident with understated warmth"] LAYER 3 — Design Language: [THE VISUAL STYLE — e.g. "editorial minimalism with a tactile human quality"] LAYER 4 — Material & Texture: [PHYSICAL QUALITIES — e.g. "natural stone, uncoated paper, soft ceramics"] LAYER 5 — Lighting & Composition: [HOW IT'S LIT AND FRAMED — e.g. "soft diffused daylight, gentle shadows, balanced composition with generous negative space"] LAYER 6 — Cultural References: [DESIGN INSPIRATIONS — e.g. "inspired by Scandinavian design and modern art galleries"] LAYER 7 — Constraints: [WHAT TO EXCLUDE — e.g. "no text, no logos, no illustrative style, no people"] LAYER 8 — Technical Parameters: [SIZE AND SETTINGS — e.g. "--ar 4:5 --style raw --v 6"] COMBINED EXAMPLE (paste this directly into Midjourney /imagine or Leonardo AI): A premium sustainable skincare brand visual world, calm and confident with understated warmth, editorial minimalism with a tactile human quality, materials: natural stone, uncoated paper, soft ceramics, soft diffused daylight with gentle shadows, balanced composition with generous negative space, inspired by Scandinavian design and modern art galleries, no text, no logos, no illustrative style --ar 4:5 --style raw --v 6 --- PROMPT 2: Asset-Specific Template (for individual deliverables) --- [ASSET TYPE + PLATFORM]: [e.g. YouTube thumbnail 1280x720 / Instagram post 1080x1080 / LinkedIn banner 1584x396] GOAL: Viewer feels [EMOTION], understands [PROMISE/MESSAGE] SUBJECT: [MAIN VISUAL ELEMENT — e.g. "confident woman at standing desk"], [ACTION], [KEY PROP] COMPOSITION: [CAMERA ANGLE — e.g. "eye-level"], [FRAMING — e.g. "medium shot"], negative space on [LEFT/RIGHT] for text overlay STYLE: brand colors [YOUR HEX CODES], [MATERIAL FROM LAYER 4], [LIGHTING FROM LAYER 5], [MOOD FROM LAYER 2] QUALITY: crisp, minimal clutter, professional, no watermarks TEXT (if needed): "[EXACT WORDS TO SHOW]" in bold sans-serif, [COLOR], [POSITION] NEGATIVE: blurry, distorted, clutter, watermark, extra fingers, deformed
Unlock this prompt
No spam. Just the prompt.
Why this works: The 8-layer architecture is a repeatable formula for any brand visual — mood boards, social posts, packaging, thumbnails. Fill in the layers once, reuse the template forever. The asset-specific prompt handles individual deliverables with exact platform sizes. Together they replace hours of design briefs and expensive Canva templates.
Batch produce 30 video scripts in a day
How to use this
- Run PROMPT 1 in ChatGPT (chatgpt.com). Fill in the fields — choose ONE option from each [OPTION A / OPTION B] list. ChatGPT generates 5 script variations.
- Pick your favorite hook from the 5 variations. Copy that script.
- Go to invideo AI (invideo.io). Click "Create new" > paste PROMPT 2 (customized with your chosen script). invideo generates a finished video with stock footage, music, and subtitles.
- Repeat 6 times with different topics. You'll have 30 scripts and 6 finished videos in one day.
--- PROMPT 1: Script Generator (run in ChatGPT) --- Generate a video script for [YOUR PLATFORM — choose one: TikTok / Instagram Reels / YouTube Shorts / LinkedIn Video]. TOPIC: [YOUR TOPIC — e.g. "3 mistakes killing your productivity"] TARGET AUDIENCE: [WHO THIS IS FOR — e.g. "startup founders who work 12-hour days"] CONSTRAINTS (choose one from each): - Duration: [15 seconds / 30 seconds / 60 seconds] - Hook style: [Pattern interrupt — shock or surprise / Curiosity gap — make them need to know / Direct benefit — tell them what they'll get] - Value format: [Single tip — one powerful idea / Listicle — 3–5 quick points / Story — personal narrative with lesson] - CTA: [YOUR DESIRED ACTION — e.g. "Follow for more" / "Link in bio" / "Save this for later"] - Tone: [Energetic and fast-paced / Professional and authoritative / Conversational and relatable] SCRIPT STRUCTURE: Hook (0–3s): Pattern interrupt that stops the scroll — no "Hey guys" or "In this video" Setup (3–10s): The problem or a controversial statement Value (10–50s): 3 actionable points (specific, not generic advice) CTA (50–60s): One clear, specific action Generate 5 VARIATIONS with different hooks. Label each variation. --- PROMPT 2: invideo AI Production (paste into invideo.io) --- Create a [DURATION]s [PLATFORM FORMAT — e.g. "Instagram Reel (9:16)"] for [YOUR AUDIENCE]. Topic: "[YOUR TOPIC]" Hook (0–2s): "[YOUR BEST HOOK FROM PROMPT 1]" Beats: [SCRIPT STRUCTURE — e.g. "3 quick mistakes with text overlays"] Close: "[YOUR CTA]" Style: punchy, quick cuts, modern stock footage, upbeat music, large readable subtitles, clean transitions.
Available when AI School launches
This prompt will be available in the app. Join the waitlist for early access.
Join the WaitlistWhy this works: ChatGPT writes 5 script variations with different hooks in seconds. Pick the best one, feed it to invideo AI, and get a finished video with stock footage, music, and subtitles. Run this 6 times with different topics and you have 30 scripts + 6 finished videos in a day.
Stop sounding like AI
ChatGPT
Grammarly·Course 57 — Stop Sounding Like AIHow to use this
- Start with any AI-generated text that sounds robotic. Paste it into ChatGPT (chatgpt.com) along with PASS 1.
- Copy the output from PASS 1. Paste it back into the same chat with PASS 2. Repeat for PASS 3 and PASS 4 — each pass fixes a different "AI tell."
- After all 4 passes, paste the final result into Grammarly (grammarly.com) for grammar and readability polish.
- The result reads like a human wrote it. Use this workflow on any AI-generated blog post, email, social post, or landing page copy.
THE 4-PASS HUMANIZER — Run these in order. After each pass, copy the output and paste it as input for the next pass. PASS 1 — Remove AI voice (paste your AI-generated text after this instruction): "Rewrite this to sound like a real person wrote it. Be direct, specific, and slightly opinionated. Remove every cliche (leverage, game-changer, at the end of the day, it's important to note). Use shorter sentences. Cut any sentence that doesn't add new information. Keep the core meaning identical." [PASTE YOUR AI-GENERATED TEXT HERE] --- After reviewing the output, copy it and start a new message: --- PASS 2 — Cut the fluff: "Make this exactly 25% shorter. Remove redundant phrases, filler sentences, and anything that restates what was already said. Every sentence must earn its place. Do not lose any key facts or arguments." [PASTE OUTPUT FROM PASS 1] --- Copy output, new message: --- PASS 3 — Warm it up: "Make this friendlier and more conversational without being cheesy. No exclamation points. No 'Great question!' energy. Think: smart friend explaining something over coffee, not a corporate email." [PASTE OUTPUT FROM PASS 2] --- Copy output, final message: --- PASS 4 — Add confidence: "Remove all hedging words: might, could, maybe, perhaps, it seems, I think, in my opinion, arguably. Replace them with direct statements. If something is uncertain, say 'This is uncertain' rather than hedging. Make every sentence decisive." [PASTE OUTPUT FROM PASS 3]
Available when AI School launches
This prompt will be available in the app. Join the waitlist for early access.
Join the WaitlistWhy this works: Each pass targets a specific AI tell: Pass 1 kills generic language, Pass 2 removes bloat, Pass 3 fixes the fake-enthusiastic tone, Pass 4 eliminates wishy-washy hedging. Four passes, four problems solved. The result reads like a human wrote it because you've systematically removed everything that screams "AI."
Send 100 personalized videos before lunch
How to use this
- Go to heygen.com. Create a new Personalized Video campaign. Select or create your AI avatar.
- Paste the script template below into HeyGen's script editor. The {{double_braces}} are merge fields that HeyGen replaces automatically for each recipient.
- Prepare a CSV file with columns: first_name, company_name, one_line_hook, cta. Fill in one row per prospect (see the CSV format below the script).
- Upload the CSV to HeyGen. Click Generate. HeyGen creates one unique video per row. Download all videos and attach them to your outreach emails for 3-5x higher reply rates.
HEYGEN SCRIPT TEMPLATE (paste into HeyGen's script editor):
Hey {{first_name}} — quick 20 seconds.
I noticed {{one_line_hook}} at {{company_name}}.
We help [YOUR INDUSTRY] teams [YOUR KEY BENEFIT — e.g. "cut ordering friction and increase upsells without slowing service"].
If useful, I can send 2 specific ideas for {{company_name}} based on what you're doing today.
{{cta}}
Either way — hope your week's going well.
---
CSV FORMAT (create this spreadsheet and upload to HeyGen):
first_name, company_name, one_line_hook, cta
Sarah, Hilton Downtown, "you just launched a new room service menu", "Want me to send those ideas? Just reply 'yes'"
Mike, The Grand Hotel, "you're expanding to 3 new locations this year", "Happy to hop on a 10-min call — here's my calendar link"
Lisa, Coastal Resort, "your team grew 40% last quarter", "I recorded a 2-min walkthrough — want me to send it?"
CUSTOMIZATION TIPS:
- {{one_line_hook}}: Reference something specific you found about them (news, LinkedIn post, job listing, product launch)
- {{cta}}: Keep it low-friction. "Reply yes" outperforms "Book a call" by 2x.
- Keep the total script under 30 seconds — shorter videos get higher completion rates.Available when AI School launches
This prompt will be available in the app. Join the waitlist for early access.
Join the WaitlistWhy this works: HeyGen's merge fields personalize an AI avatar video for every prospect — their name, their company, a specific observation about their business. Record one template, upload a CSV of 100 prospects, and HeyGen generates 100 unique videos. Sales teams report 3-5x higher reply rates vs. text-only outreach.
Featured Courses
Start your content creator plan and master these first.
Get notified when AI School launches.
Be the first to access 66+ courses on 50+ AI tools.
Join the WaitlistFrequently Asked Questions
Can AI create videos without me being on camera?
Yes. Synthesia lets you create professional training videos and content using AI avatars. You write the script, choose an avatar and voice, and it produces broadcast-quality video — no camera, no studio, no editing. One prompt can turn a document into a 3-video training series.
How do I create brand visuals with AI?
Use Midjourney or Leonardo AI with structured prompts that define subject, emotional tone, design language, materials, lighting, composition, and cultural references. The 8-layer prompt architecture produces consistent, brand-aligned visuals for mood boards, thumbnails, and marketing assets.
Can AI batch-produce video scripts?
Absolutely. With a structured prompt that defines platform, duration, hook style, value format, and CTA, AI generates 5 script variations in seconds. Feed the best ones into invideo AI to produce finished short-form videos with stock footage, music, and subtitles.
How do I make AI-generated content sound human?
Use a 4-prompt humanization toolkit: (1) rewrite for directness and specificity, (2) cut 25% of the length, (3) warm up the tone without exclamation points, (4) remove hedging words like 'might' and 'could'. Pair with Grammarly for a final polish.
What AI tools do content creators use?
Synthesia and HeyGen for video without cameras, Midjourney and Leonardo AI for visuals, ChatGPT for scripts and copy, invideo AI for short-form video production, and Grammarly for humanizing AI-generated text.







