9 ideas on how to get recommended by Google's "Personal Intelligence"

Make your receipt emails sell for you
Your order confirmation and shipping emails are now part of your brand footprint. Make the product name and variant painfully clear, add a short “how to use it” section, and include a relevant cross-sell that feels helpful. If AI is using inbox context, you want your receipts to be clean and unmistakable.

Create a “first purchase” product that’s an easy yes
You need a gateway product. Something low risk, high satisfaction, and easy to choose. If personalised recommendations reward brands people bought from before, you want more people buying once, sooner.

Add “use-case blocks” to PDPs (weather, occasions, fit, care)
AI recommendations will reward products that are easy to match to a real situation. Add a simple block that answers: when to wear it, what weather it suits, how it fits, how to care for it, and what it pairs well with. This helps customers decide faster and helps machines understand what the product is actually for.

Build “occasion bundles”
Personalised search will be context-heavy. So make bundles that match context:
- Holiday packing bundle
- Office essentials bundle
- Wedding guest bundle
- Gym starter kit

Build comparison pages
Create pages to get recommended with when you competitors get mentioned, pages like:
- “Brand A vs Brand B”
- “Is this worth it vs cheaper options?”
- “What to buy instead of X”
Include key differentiators and reasons why your brand is superior.

Turn returns into a conversion weapon
If customers are returning competitors’ stuff, they’ll be desperate for clarity. Over-index on fit guides, “fits like” notes, and real customer photos. Reduce regret. Become the brand the AI treats as the “buy it and keep it” option.

Make your product naming and variant data consistent
This is not sexy, but it’s a massive edge. Same product name across your site, feeds, and emails. Same variant naming everywhere. The brand with clean data becomes easier for the AI to understand, trust, and recommend.

Use “keepable” emails to stay in the customer’s inbox
Send emails people won’t delete: care guides, styling ideas, reminders, warranty info, reorder timing. If inbox context is part of the recommendation loop, these emails become signals that reinforce your brand and products over time.

Use abandoned cart emails
Your abandoned cart emails should be a clean, useful “memory marker” that AI can latch onto. Clear product name, clear variant, clear delivery promise, clear returns, and one simple reason to come back. If the AI is using inbox context, you want it to see a strong, consistent trail that says: this person nearly bought this, from this brand, at this price.