Label your AI product images

Hey there Smarty 👋

Are you geared up to catch the latest and greatest in quick shorts?

And just a quick heads-up! If you stumbled upon us through a friend, make sure to subscribe here! That way, you’ll never miss out on the trending shorts.

🖼️ AI made your product shots. google wants that on the record

Plenty of catalogs now carry AI-generated lifestyle scenes, models and descriptions. Google's AI content guidance, refreshed on October 1, is specific about ecommerce.

AI-generated images should carry IPTC metadata marking them as made by an algorithm, and AI-generated product data should be specified separately and labeled. Better to sort that out before Q4 traffic peaks.

Inventory what AI touched

Open a sheet with one row per product and four columns:

1. Main image: photographed, AI-edited or fully generated
2. Lifestyle images: same three options
3. Description: human-written, AI-drafted and edited, or AI-generated
4. Specs: confirmed against supplier data or not

In my view, those specs matter beyond compliance, because they're the raw material AI shopping answers work from. That gap between being recommended and merely cited is the focus on October 15, when AirOps COO Matt Hammel and Bain & Company's Zak Prauer cover how to invest in AI search before attribution catches up. It's free, with recordings within 24 hours. You can secure your free spot.

When in doubt, treat a fully generated scene or model as AI-generated. Background cleanup on a real photo is a gray area, so flag it for review rather than guessing.

A home decor brand might find its product shots are real but every room scene behind them was generated. That's the column to fix first.

Write the metadata into the files

For every image marked AI-generated, set the IPTC Digital Source Type to "trainedAlgorithmicMedia" before upload.

With the free ExifTool utility, one command handles a whole folder:

exiftool -XMP-iptcExt:DigitalSourceType="http://cv.iptc.org/newscodes/digitalsourcetype/trainedAlgorithmicMedia" ./ai-scenes/

Then add a check to your upload process: no AI-scene file goes live without that field set.

Once a week, spot-check five live images with exiftool -DigitalSourceType in case your image pipeline strips metadata on upload. If the field is gone, fix the pipeline before re-uploading.

Keep verified specs apart from AI copy

Split each product page into two layers:

• Spec block: dimensions, materials, weight and care. Every value is sourced from supplier or lab data, with no AI drafting.
• Description: AI-assisted is fine here, but a human edits it and it never states a spec that isn't in the block above.

If an AI scene shows the product doing something it can't, such as a jacket in a downpour when it isn't waterproof, swap the image. Misleading visuals tend to come back as returns.

If a supplier can't confirm a spec, pull it from both layers until they do. Fill in the inventory for your top 50 products before Friday.

Together with Omnisend

Still Hand-Holding AI Through Your BFCM Prep?

If you’re still copying reports into AI and explaining what they mean, the time savings disappear before you’ve made a single campaign decision.

Omnisend’s free BFCM AI playbook gives you twelve detailed prompts to get answers from your store’s data without starting every question from scratch.

Find last year’s offers worth repeating and buyers who haven’t returned. Check for deliverability warning signs before you increase sending volume, while there’s still time to address them.

Omnisend connects with Claude or ChatGPT, so you can analyze performance and create campaigns in the same conversation. Put those findings into a campaign calendar you can actually work from.

New to Omnisend? Its email and SMS platform serves 150,000+ ecommerce brands, and free migration takes up to five business days.

Start while there’s time to use what you learn. You’ll have more room to test your campaigns before the BFCM rush.

Try Omnisend now.

🎯 Meta, X, Reddit and Pinterest are making ads easier to launch

Four major advertising platforms are rolling out updates that simplify campaign creation, improve targeting and open new ways to promote content, with AI handling more of the manual work.

The Breakdown:

Meta Brings AI Into Ads Manager - Meta is testing an AI assistant that builds campaigns through chat, adjusts targeting and budgets, and remembers past tests and seasonal trends.

Meta Connects Creative to Checkout - Advantage+ can now turn static images into videos, while upcoming Messenger and Meta AI checkout tests will let shoppers purchase directly within conversations.

X Builds Ads From a URL - X Lift generates ad creative and targeting from a website link, leaving advertisers to choose creatives and budgets. X reports 50% annual growth in active self-serve advertisers.

Reddit Targets Buyers More Precisely - Reddit is testing B2B in-market audiences across enterprise, business AI and cybersecurity. Early testers saw nearly 7% lower CPA, while Max campaigns now support lead generation.

Pinterest Opens Creator Promotions - Brands can now boost existing creator Pins without commissioning the content. Creators receive an email and can decline or opt out of promotions entirely.

Ad platforms are removing more manual steps between planning and launching campaigns. AI can increasingly build ads, adjust targeting and simplify buying, but advertisers still need to decide which audiences, creatives and outcomes are worth spending money on.

👨‍💻 Quick Hits

📺 TikTok and Kick say they would prohibit a proposed livestream of a U.S. military execution, while YouTube, Meta, Twitch and X have not publicly clarified whether they would allow the broadcast.

🔓 Meta has removed separate access verification for developers accessing clients’ business data, allowing apps to go live faster. Meta says its existing Business Verification and App Review processes provide sufficient safeguards.

🛠️ X has rebuilt its API Exhibit website, showcasing developer apps, integrations and agentic use cases. Developers can now customize listings and submit Grok Bot templates and agent skill files to increase visibility.

🔍 Google AI Overviews appeared in 82.9% of U.S. desktop branded-search results in Ahrefs’ September 29 tracking, while a separate 100-brand study found 63% triggered AI Overviews, mostly appearing in the top position.

🥳 Events 

🔥 Why Google May Be Getting Credit for Customers AI Helped Win

October 15 | Virtual Event

A buyer can discover your brand in AI, convert through Google, and leave your attribution report telling the wrong story. Matt Hammel and Bain’s Zak Prauer show how to evaluate recommendation influence, spot what paid-channel metrics miss, and make a stronger case for AI search investment.

✅ Reserve Your Free Spot

Can't attend the session live? Register anyway, and you’ll get the recordings in 24 hours.

That’s a wrap for today! Tell us your thoughts about today's content as we line up more Shorts! And don’t hesitate to share this with someone who’d adore it. 🥰