Amazon and YouTube race to make AI your creative partner for ads and Shorts
Amazon folds agentic AI into Seller Assistant for ad creation and planning. YouTube adds text-to-video Shorts, speeding production and tightening the loop from idea to publish.

Amazon and YouTube just rewired how ads and videos get made
Two giants made major AI moves in September 2025. Amazon folded agentic AI into its advertising stack, and YouTube opened up text-to-video for Shorts. For creatives, that means faster production, smarter planning, and fewer blockers between idea and output.
What Amazon shipped
On September 18, Amazon integrated video, audio, and image generation from Creative Studio directly into its ad platform. These tools sit inside Seller Assistant, so you can plan and produce ads through a simple chat flow.
- Available free for all sellers, rolling out over the next few months.
- Create professional ad assets from conversational prompts in hours instead of weeks.
- AI analyzes your products and Amazon shopping signals, then explains why it made certain creative and targeting choices.
It goes beyond creative. Seller Assistant reviews sales patterns, proposes new product variants, drafts growth plans, adjusts inventory, flags slow movers, coordinates drop-shipping, and surfaces compliance requirements by region. Even documentation for complex categories-like electrical-can be prepared by the system.
Proof it works: Bird Buddy used Creative Studio to build a Sponsored Video for Father's Day and saw a 338% higher click-through rate than their other Sponsored Video ads, 89% new-to-brand offers, and a 121% return on ad spend. The video had minor quirks (like an odd bird foot), but the performance lift was clear.
Amazon's ad business is already up 23% year-over-year with annualized revenue between A$90 and $100 billion. The company says its agentic AI-AI that reasons, plans, and takes action-will keep getting smarter as it learns from seller needs.
What YouTube rolled out
On September 16, YouTube introduced new AI features with a focus on creator control. Headlining the update: a Shorts tool powered by a custom version of Google DeepMind's Veo 3 that turns a text prompt into a vertical video with sound-no camera or editing suite needed. It's rolling out in the US, UK, Canada, Australia, and New Zealand first.
- Edit with AI to turn raw footage into a first draft.
- Speech to Song to convert dialogue from eligible videos into music.
- Experimental Veo features that add motion and objects to Shorts.
- Expanded facial likeness detection in YouTube Studio (open beta for YouTube Partner Program creators) to request removal of unauthorized videos.
- Ask Studio, an AI chat tool for Studio (US only at launch).
- Collaborators feature so a video appears to all creators' audiences.
- A/B testing for titles, Inspiration Tab updates, and auto-dubbing with lip sync in limited testing.
As YouTube's CEO put it, the goal is AI that serves human creation-not the other way around.
Why this matters for creatives
- Fewer bottlenecks: Concept, production, and iteration can happen in a single session.
- Better decisions: Built-in analysis guides creative choices, targeting, and launch timing.
- Cost control: Small teams can compete with bigger budgets using automated asset creation.
- Brand safety: New tools help detect and remove unauthorized likeness use.
Practical playbook for the next 30 days
- Ship three concepts: Use Amazon's Seller Assistant to generate three distinct ad creatives for one product. A/B test headlines, visual style, and CTA. Set a 7-day budget cap and compare CTR and new-to-brand rates.
- Shorts sprint: Draft five Shorts with Veo 3 prompts. Keep scripts under 45 seconds. Publish on a fixed cadence and test topic clusters (product demo, behind-the-scenes, community story, myth-busting, FAQ).
- Guardrails: Create a brand style prompt (tone, pacing, color, typography, visual motifs). Paste it into each session to keep outputs consistent.
- QA checklist: Before publishing, scan for visual artifacts, off-brand phrasing, and compliance flags. Fix in post or regenerate.
- Measure weekly: Track CTR, watch time, completion rate, and ROAS. Keep what works. Kill what doesn't. Iterate fast.
Prompt templates you can copy
- Amazon Seller Assistant - performance ad
"You are a senior ad strategist. Product: [product]. Audience: [who they are]. Goal: [e.g., increase new-to-brand purchases]. Constraints: [brand tone, visual style, budget]. Produce: 3 campaign concepts with headlines, 15-sec video scripts, image variants, and targeting suggestions. Explain why each choice fits the audience and goal." - YouTube Shorts - text-to-video
"Create a 30-45 sec vertical video that shows [core scene] with [style reference: moody, high-contrast, macro details]. Include [on-screen text beats], [sound/mood], and [closing CTA]. Keep pacing tight and the hook in the first 2 seconds."
Risks and how to handle them
- Visual oddities: Always preview. If an asset looks off, regenerate or switch styles.
- Brand drift: Lock a short, reusable brand brief. Reuse it in every prompt.
- Likeness misuse: Use YouTube's facial likeness tool to request removals if needed.
- Over-automation: Keep a human editor in the loop for final review and compliance.
Where to explore more
- CNET coverage of YouTube's AI features
- Mumbrella reporting on Amazon's ad updates
- Generative video tools directory - Complete AI Training
The bottom line: AI is now a creative partner inside the platforms where your work ships. Use it to produce more, test more, and learn faster-while you keep taste and judgment firmly in charge.