Veo 3.1 “Ad Read” Prompt Pack (2026): 12 Creator-Ready Templates for UGC, Product Demos, and Service Offers (with a 3-Step Rewrite Method)

Veo 3.1 “Ad Read” Prompt Pack (2026): 12 Creator-Ready Templates for UGC, Product Demos, and Service Offers (with a 3-Step Rewrite Method)

If you’ve ever prompted an AI video model for an “ad” and got something that looks cinematic—but sells nothing—this is for you.

This post is a practical mini-library of Veo 3.1 ad prompts you can copy/paste, plus a simple way to convert any existing script into shot cards that generate more reliably.

All guidance below reflects publicly available prompting advice as of 2026-04-24, including Veo prompting guidance that encourages specifying camera/framing/motion, lighting, style references, detailed character/location descriptions, and even dialogue direction when needed. (https://deepmind.google/models/veo/prompt-guide/) Veo 3.1 is also described as generally available for production on Vertex AI as of 2026-04-24. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

What makes an AI-generated ad “work” (and what usually breaks)

Most “bad AI ads” aren’t failing because of creativity—they fail because the prompt doesn’t control the basics:

• Unclear subject: The model doesn’t know what the product is, who’s holding it, or what to show.
• Too many actions at once: One shot tries to unbox, explain features, show results, and end with a CTA.
• Vague visual direction: “Modern vibe” is less useful than camera, lighting, location, and style texture.

Veo prompting guidance consistently rewards specificity—especially around shot framing/motion, lighting, style/texture, location detail, and character detail/action. (https://deepmind.google/models/veo/prompt-guide/)

The 3-step rewrite method: Script → Hook → Shot Cards

Instead of prompting “make me a 30s ad,” rewrite into a structure Veo can follow.

Step 1) Script: reduce to one promise + one proof + one CTA

You’re aiming for a single spine:

• Promise: what changes for the viewer
• Proof: why they should believe it (demo moment, metric, testimonial line)
• CTA: what to do next

Step 2) Hook: pick one of these 3 variants (swap without changing the rest)

Use one of these hook styles at the top of any template below:

• Curiosity hook: “I didn’t expect {product} to fix {pain}… but here’s what happened.”
• Pain-first hook: “If you’re still dealing with {pain}, this is for you.”
• Results-first hook: “I got {result} in {time} using {product}.”

Step 3) Shot Cards: split 20–30s into 3–6 shots

One action per shot keeps generations cleaner.

• Shot 1: Hook (face + product glimpse)
• Shot 2: Problem context (one clear visual)
• Shot 3: Demo (one feature, one action)
• Shot 4: Proof moment (result, comparison, social proof line)
• Shot 5: Offer + CTA (clean framing, minimal text)

Veo prompting guidance notes that for fast-paced scenes, you can map out exact play-by-plays with highly detailed prompts. (https://deepmind.google/models/veo/prompt-guide/)

Prompt templates #1–4: UGC-style ads

How to use: pick a hook variant, fill the slots, and keep on-screen text short.

Template 1) UGC Testimonial “Ad Read”

Copy/paste prompt:

Create a 20–25s vertical UGC ad for {product} targeting {audience}. Brand style: {brand style}.
Hook (choose one): {HOOK_CURIOSITY | HOOK_PAIN_FIRST | HOOK_RESULTS_FIRST}.
Key message: solve {pain} with the promise {promise}. Include proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shot cards (one action per shot, 4–5 shots):

1. Subject: relatable creator (describe appearance, voice, and vibe in detail), holding {product}. Action: speaks hook to camera. Setting: bright kitchen or bedroom. Camera: medium close-up, slight handheld, gentle push-in. Lighting: warm, even. Style: natural smartphone look. On-screen text: 3–5 words summarizing hook, large and legible.
2. Action: creator shows the “before” pain moment related to {pain}. Setting: same room. Camera: over-the-shoulder or close-up on problem. Lighting consistent. On-screen text: one short pain phrase.
3. Action: creator uses {product} (single clear step). Camera: close-up of hands/product, slow pan. On-screen text: one benefit word.
4. Action: creator states proof {proof} and delivers the promise {promise}. Camera: back to face, confident. On-screen text: one proof keyword.
5. Action: creator says offer {offer} and CTA {CTA}. Camera: steady, centered. On-screen text: CTA only.
   Audio: clear spoken dialogue matching the script; subtle room tone.

(Notes: Veo prompting guidance supports specifying camera motion/framing, lighting, style references, and detailed character description and dialogue direction. (https://deepmind.google/models/veo/prompt-guide/))

Template 2) Problem → Solution “Talk to Camera”

Generate a 20–30s vertical direct-response ad for {product}. Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Then: “I used to struggle with {pain}. Now I get {promise}.” Include proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (3–4 shots):

1. Subject: creator, expressive, direct eye contact. Action: hook line. Setting: neutral wall. Camera: tight close-up, slight low-angle. Lighting: soft key light. Style: clean UGC. Text: hook phrase only.
2. Action: visualize {pain} with one simple prop/action. Camera: close-up, quick tilt. Lighting same. Text: “{pain}” (2–3 words).
3. Action: show {product} solving it (single step). Camera: top-down or close-up, slow pan across product. Text: “{promise}” (max 4 words).
4. Action: creator closes with offer + CTA. Camera: medium close-up, stable. Text: CTA only.

Template 3) Unboxing → First Use → Reaction

Create a 25–30s vertical unboxing ad for {product} for {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Include promise {promise}, proof {proof}, offer {offer}, CTA {CTA}.

Shot cards (5–6 shots):

1. Subject: creator at table. Action: holds sealed package, hook line. Camera: medium shot, slight handheld. Lighting: bright, even. Text: hook (3–5 words).
2. Action: unbox (one continuous action). Camera: top-down, slow push-in. Text: “Unboxing {product}”.
3. Action: show “hero” product close-up. Camera: macro/close-up, slow pan. Style/texture: crisp, high detail. Text: 1 feature word.
4. Action: first use (single step). Camera: close-up on hands. Text: “Step 1”.
5. Action: reaction + proof {proof}. Camera: face close-up, authentic. Text: proof keyword.
6. Action: offer + CTA. Camera: centered, stable. Text: CTA only.

Template 4) “Day in the Life” (Soft-sell UGC)

Create a 25–30s vertical day-in-the-life ad featuring {product}. Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Theme: show {product} naturally solving {pain} and delivering {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (4–6 shots, quick cuts, one action per shot): morning setup → problem moment → product use → result moment → CTA.
For each shot specify: setting details, camera framing, and lighting (warm morning light, etc.). Keep on-screen text minimal, large, high-contrast.

Prompt templates #5–8: Product & feature demos

These are for when you need clarity over charisma.

Template 5) “How it works” 3-step demo

Create a 20–25s vertical product demo ad for {product} for {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Promise: {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

4 shots:

1. Hook on face + product. Camera: medium close-up, gentle push-in. Lighting: bright, even. Text: hook phrase.
2. Step 1: {product} setup. Camera: top-down, steady. Text: “1) {Step1_label}”.
3. Step 2: core action (single feature). Camera: close-up, slow pan. Text: “2) {Step2_label}”.
4. Step 3 + result: show outcome + proof {proof}. Camera: split-focus or before/after cut. Text: “3) {Step3_label}” then CTA.

(Using a layered prompt structure like camera/subject/action/environment/lighting/style/audio is commonly presented as a repeatable formula for Veo 3.1 prompting. (https://invideo.io/blog/google-veo-prompt-guide/))

Template 6) Before/After (single variable)

Generate a 20s vertical before/after ad for {product}. Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Pain: {pain}. Promise: {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (3–5):

1. BEFORE: show {pain} clearly (one visual). Camera: close-up. Lighting: neutral. Text: “Before”.
2. Demo: show {product} used (one step). Camera: close-up on product. Lighting: brighter. Text: 1 benefit word.
3. AFTER: show result aligned with {promise}. Camera: matching framing to shot 1 for clarity. Text: “After”.
4. Proof moment: {proof} stated on camera. Text: proof keyword.
5. CTA: offer + CTA, clean background. Text: CTA only.

Template 7) Comparison (A vs B)

Create a 25–30s vertical comparison ad. Compare {product} vs {alternative}. Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Focus on one comparison point: {pain} → {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (4–5):

1. Hook on face. Text: hook.
2. Show {alternative} problem moment (single issue). Camera: close-up. Text: “Old way”.
3. Show {product} solution moment (single step). Camera: close-up, slow pan across product. Text: “New way”.
4. Proof: {proof}. Camera: medium close-up. Text: proof keyword.
5. CTA: offer + CTA. Text: CTA only.

Template 8) Checklist demo (“3 things to look for…”)

Create a 20–25s vertical checklist ad for {product}. Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Pain: {pain}. Promise: {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (4):

1. Hook to camera. Text: “3 things to check”.
2. Checklist item #1 shown with {product}. Camera: close-up. Text: “1) {Item1}”.
3. Checklist item #2 shown. Camera: close-up. Text: “2) {Item2}”.
4. Checklist item #3 + proof + CTA. Camera: medium close-up, stable. Text: “3) {Item3}” then CTA.

Prompt templates #9–12: Service offers

Template 9) Local business “Book now”

Create a 20–25s vertical local service ad for {product} (service). Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Pain: {pain}. Promise: {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (4–5): storefront/clinic exterior → service in action (one procedure step) → customer reaction → proof line → booking CTA.
Camera: include one establishing wide shot, then close-ups. Lighting: clean, welcoming. On-screen text: location + CTA only.

Template 10) SaaS screen-focused demo (simple UI moments)

Generate a 25–30s vertical SaaS ad for {product}. Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Pain: {pain}. Promise: {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (5):

1. Creator hook to camera.
2. Screen moment: show one UI action that addresses {pain}. Camera: over-the-shoulder laptop shot. Text: “Step 1”.
3. Screen moment: show the key feature. Text: “Step 2”.
4. Result screen: show outcome aligned with {promise}. Text: proof keyword.
5. CTA: offer + CTA on clean background.

Template 11) Lead magnet “Free guide”

Create a 20–25s vertical lead magnet ad for {product} (free resource). Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Pain: {pain}. Promise: {promise}. Proof: {proof}. Offer: {offer}. CTA: {CTA}.

Shots (4): hook → show the guide/resource (close-up) → preview 2–3 bullet benefits (fast cuts) → CTA to download. Keep text minimal and readable.

Template 12) Webinar/event registration

Create a 25–30s vertical event ad for {product} (webinar/event). Audience: {audience}. Brand style: {brand style}.
Hook: {HOOK_*}. Pain: {pain}. Promise: {promise}. Proof: {proof}. Offer: {offer} (include date/time placeholders). CTA: {CTA}.

Shots (5): host hook → event title card (minimal text) → 2 quick proof moments (credentials or outcomes) → agenda tease (3 bullets max) → CTA to register.

Add-ons that boost conversion clarity (without overstuffing)

Keep on-screen text minimal and legible

Treat text like subtitles, not a paragraph. Use:

• 3–5 words per line
• 1–2 lines max
• high contrast
• safe margins

Time the CTA twice

• Soft CTA mid-way (“link in bio” / “tap to learn more”)
• Hard CTA in final 2–3 seconds

Add a “proof moment” shot

Proof can be a quick demo result, a testimonial line, or a single metric—keep it visual and singular.

Common issues + fast fixes (prompt rewrites)

Symptom	Likely cause	Prompt rewrite
Product barely appears	Subject/action not explicit	Add: “{product} must be visible in shots 1, 3, and final CTA shot; include a close-up hero shot.”
CTA gets ignored	CTA only mentioned once, late	Add: “Include a mid-ad soft CTA + final CTA; final shot must be clean background with CTA text only.”
Messy or tiny on-screen text	Too much text; no constraints	Add: “On-screen text max 5 words, large, high-contrast, one line; avoid dense paragraphs.”
Off-brand tone	{brand style} too vague	Add style references and texture cues (e.g., “film noir,” “slightly worn VHS texture”) and specify voice/vibe. (https://deepmind.google/models/veo/prompt-guide/)
Scene looks generic	Location not described	Add sensory, specific location details (time of day, atmosphere, materials). (https://deepmind.google/models/veo/prompt-guide/)
Actor feels inconsistent	Character description too brief	Provide detailed appearance/voice/action notes per shot. (https://deepmind.google/models/veo/prompt-guide/)

Copy/paste: One-page “Ad Prompt Brief” (fill this in first)

Use this as your preflight so every prompt has the same backbone.

Ad Prompt Brief (fill-in):

• {product}:
• {audience}:
• {pain}:
• {promise}:
• {proof}:
• {offer}:
• {CTA}:
• {brand style}:
• Aspect ratio: 9:16 (vertical) / 1:1 / 16:9
• Required shots (3–6):
• Must-show visuals (product close-up, before/after, UI moment, etc.):
• On-screen text rules (max words, placement):

Quick checklist (before you generate)

• Hook chosen (curiosity / pain-first / results-first)
• 3–6 shots, one action per shot
• Product clearly visible in at least 2 shots
• On-screen text is short, large, and high-contrast
• CTA appears in the final shot (clean frame)

FAQ

Can Veo handle dialogue for ad reads?

Veo prompting guidance indicates the model can generate dialogue, and you can either set a topic or provide specific lines to say. (https://deepmind.google/models/veo/prompt-guide/)

What prompt details matter most for ad reliability?

Public guidance emphasizes specifying camera framing/motion, lighting, style/texture references, and detailed character and location descriptions to get more specific results. (https://deepmind.google/models/veo/prompt-guide/)

How long should I make each shot?

For a 20–30s ad, 3–6 shots is a practical range; the key is making each shot do one job (hook, demo, proof, CTA) rather than cramming everything into one generation.

Does this guarantee better ROAS?

No—these templates are about production clarity and iteration speed, not guaranteed performance outcomes.

Related reading

• Veo 3 API pricing comparison
• Veo 3.1 vs Sora 2 comparison
• Getting started with the Veo 3 API

CTA: Generate faster with Veo3Gen

If you want to turn these shot cards into repeatable workflows (and iterate variations cleanly), explore the Veo3Gen endpoints and plans:

• Build with the API: /api
• See plans and limits: /pricing

Try Veo3Gen (Affordable Veo 3.1 Access)

If you want to turn these tips into real clips today, try Veo3Gen:

• Start generating via the API: /api
• See plans and pricing: /pricing