How to Create Marketing Videos with AI: Step-by-Step Guide Using Synthesia, HeyGen & CapCut (2026)

Step 1: Set Up Your Script

Before touching Synthesia, write your script. This is where 80% of your video quality comes from. Use this ecommerce product demo structure:

1. Hook (5–10 seconds): State the problem your product solves. “Tired of spending 3 hours editing product photos? Here’s a faster way.”
2. Introduction (10–15 seconds): Name the product, one-sentence value proposition.
3. Feature Walkthrough (60–90 seconds): Cover 3–4 key features. One feature per scene. Lead with the benefit, then show the feature.
4. Social Proof (15 seconds): One customer quote or stat. “Over 10,000 sellers use this daily.”
5. CTA (10 seconds): Clear next step. “Click below to start your free trial.”

Pro tip: Keep sentences under 20 words. AI voices handle short, punchy sentences much better than long, complex ones. Write like you speak, not like you write.

Step 2: Create a New Video in Synthesia

Log into app.synthesia.io and click “Create Video.”

1. Select “Start from blank” or choose a template. For product demos, the “Product Showcase” template saves time.
2. Choose your aspect ratio: 16:9 for website/YouTube, 9:16 for social stories, 1:1 for Instagram feed.
3. Select an AI avatar. For ecommerce, avatars like “Anna” or “James” work well — professional but approachable. If you’re on the Enterprise plan, upload a photo or video to create a custom avatar that matches your brand.
4. Choose a voice. Match it to your audience: US English for American customers, British English for UK markets. Synthesia’s latest voices (marked “Enhanced”) sound significantly more natural.

Step 3: Build Your Scenes

Each scene should correspond to one section of your script:

1. Scene 1 — Hook: Avatar on screen with a bold text overlay stating the problem. Background: solid brand color or lifestyle image.
2. Scene 2 — Introduction: Avatar introduces the product. Add your product image or logo as a screen element. Synthesia lets you position it next to the avatar.
3. Scenes 3–5 — Features: For each feature, paste your script into the text box. Upload a product screenshot or photo for each feature. Use Synthesia’s screen recording embed if you’re demoing software.
4. Scene 6 — Social proof: Text overlay with customer quote. Avatar reads the testimonial context.
5. Scene 7 — CTA: Avatar delivers the call to action. Add a clickable button overlay (available on paid plans) linking to your product page.

For each scene, paste the relevant script portion into the Script field. The avatar will lip-sync automatically.

Step 4: Add Branding & Polish

1. Upload your brand kit (logo, colors, fonts) under Settings → Brand Kit. Synthesia will apply them across all scenes.
2. Add background music from Synthesia’s royalty-free library. Keep it subtle — ecommerce product videos need the voice to be crystal clear.
3. Add captions: Turn on auto-captions. 85% of social video is watched on mute, and captions boost engagement even on product pages.
4. Review each scene with the Preview button. Check pacing — if the avatar speaks too fast, break long sentences into two shorter ones.

Step 5: Generate & Export

1. Click “Generate Video”. Synthesia typically processes a 2-minute video in 10–15 minutes.
2. Once ready, download in 1080p MP4 for your website, or use the embed link for direct page embedding.
3. For Shopify/WooCommerce product pages, embed using an HTML block or upload to YouTube/Vimeo and embed the player.

Total time: ~45 minutes for your first video. Once you have templates, subsequent videos take 15–20 minutes.

Step 1: Plan Your UGC-Style Script

UGC ads follow a different structure than polished demos. The key is authenticity over polish:

1. Hook (3–5 seconds): Start with a pattern interrupt. “I was skeptical about this product until…” or “This is the product that finally fixed my [problem].”
2. Problem (10 seconds): Relate to the viewer’s pain point.
3. Discovery (10 seconds): “I found this [product] and decided to try it.”
4. Experience (20 seconds): Walk through 2–3 specific benefits with concrete results. Numbers sell: “In two weeks, I saw a 40% improvement.”
5. CTA (5–10 seconds): “Link in bio” or “Use code X for 20% off.”

Script length: For a 60-second video, aim for 140–160 words. AI voices speak at roughly 150 words per minute.

Step 2: Choose Your Avatar & Style

Log into app.heygen.com and click “Create Video” → “Avatar Video.”

1. Select an avatar that matches your target audience. For UGC-style content, pick avatars labeled “Casual” or “Lifestyle” — they use more natural gestures and expressions.
2. Pick the setting: HeyGen lets you choose different backgrounds. For UGC-style, use a home/lifestyle setting (kitchen, living room) rather than a corporate background.
3. Photo Avatar option: Upload a single photo of a real person (with consent) and HeyGen will animate it. This creates an incredibly realistic “real person” feel at a fraction of the cost of hiring a creator.
4. Video Translate feature: If you already have a customer testimonial in English, use HeyGen’s Video Translate to create versions in Spanish, French, German, or any of 40+ languages — with lip-sync. Massive for international ecommerce.

Step 3: Build the Video

1. Paste your script into the editor. HeyGen processes one scene at a time.
2. Add product B-roll: Upload product photos or short clips. HeyGen lets you insert them as picture-in-picture overlays or full-screen cutaways between avatar segments.
3. Add text overlays: For UGC ads, add bold captions for key phrases — “40% improvement in 2 weeks” — as on-screen text. This reinforces the message for muted viewers.
4. Adjust voice settings: Set the speed slightly slower (0.9x) for a more natural, conversational pace. Select voice emotion if available — “friendly” or “enthusiastic” works best for product promotions.
5. Add background music: HeyGen includes royalty-free tracks. For UGC-style, choose something subtle and trending — lo-fi beats or acoustic guitar work well.

Step 4: Personalization at Scale (Optional)

This is where HeyGen shines for ecommerce. Use the Personalized Video API to generate unique videos for individual customers:

1. Create a template video with variable fields: “{customer_name}”, “{product_name}”, “{discount_code}”.
2. Connect via HeyGen’s API to your email platform (Klaviyo, Mailchimp) or Shopify flow.
3. Automatically generate personalized videos for abandoned cart emails, post-purchase thank-yous, or VIP customer outreach.

Brands using personalized video in abandoned cart sequences report 2–4x higher click-through rates compared to standard text emails.

Step 5: Export for Ads

1. Click “Submit” to render. HeyGen processes videos in 5–10 minutes.
2. Download in 9:16 for TikTok/Reels or 1:1 for Meta feed ads.
3. Pro tip for Meta Ads: Export the same script in 3 different avatars. Run them as A/B/C creative variants. We’ve seen a 15–25% variance in performance between avatars — testing is essential.

Total time: ~30 minutes per video. With templates, batch-produce 5+ variants in an hour.

Step 1: Gather Your Raw Materials

CapCut’s strength is transforming existing assets into social-ready content. Before starting, collect:

• 5–10 product photos (high quality, different angles)
• Any raw video clips (unboxing, product in use, customer reviews)
• Your product’s key selling points (3–5 bullet points)
• Trending audio reference (check TikTok’s trending sounds)

Step 2: Use CapCut’s AI Script Generator

Open CapCut (desktop or web at capcut.com) and start a new project.

1. Click “AI Script” in the toolbar (or navigate to the Script tab).
2. Enter your product name, category, and 2–3 key benefits.
3. Select tone: “Energetic,” “Casual,” or “Professional.”
4. Select format: “Product showcase,” “Before/After,” or “Listicle.”
5. CapCut generates a script with scene-by-scene directions and suggested timing. You can regenerate or edit manually.

Why this matters: The AI script generator understands TikTok/Reels pacing. It front-loads the hook, keeps individual scenes to 3–5 seconds, and builds toward a CTA — all optimized for the social algorithm’s retention metrics.

Step 3: Apply a Magic Template or Build from Scratch

Option A — Magic Template (fastest):

1. Browse CapCut’s template library → filter by “E-commerce” or “Product”.
2. Select a trending template. CapCut shows usage stats so you can pick formats that are currently performing well.
3. Swap in your product photos/clips where the template indicates. The transitions, timing, and effects are pre-built.
4. Edit the text overlays with your product name, price, and CTA.

Option B — Custom Build:

1. Set aspect ratio to 9:16.
2. Import your product photos and clips to the timeline.
3. Use AI Background Remover on product photos for clean, floating-product effects.
4. Apply CapCut transitions — the “Zoom Blur” and “Glitch” transitions are trending for product reveals.
5. Add text animations for prices, features, and CTAs. CapCut’s “Typewriter” and “Pop” effects work well for ecommerce.

Step 4: Add AI Voiceover & Auto-Captions

1. Click “Text to Speech” and paste your script. Choose from CapCut’s AI voices — the “Jessie” and “Male Narrator” voices are the most natural-sounding for product content.
2. Adjust pacing to match your visual cuts. The voiceover should align with scene transitions.
3. Enable “Auto Captions”: CapCut auto-generates word-by-word animated captions. Select a caption style — bold, outline, or highlighted. These are essential for social video performance.
4. Add trending music from CapCut’s licensed library. Lower the music volume to 15–20% so the voiceover stays clear.

Step 5: Export & Publish

1. Export at 1080p or 4K (4K available on Pro plan).
2. For TikTok: Export and upload directly via CapCut’s TikTok integration (one-click publish).
3. For Instagram Reels: Download and upload manually, or use Meta Business Suite.
4. For Meta Ads: Export as MP4, then upload to Ads Manager as a creative asset.

Total time: 15–20 minutes with a template, 30–40 minutes for custom builds. For high-volume stores, batch 5–10 videos in a single session.