Personalize content

Personalization changes what each customer sees. Marketing agent builds it from your customer and product data for email and SMS. Content variants, which tailor messages to audience groups, have their own page.

Types of personalization

Type	What it does	Example
Content variants	Show each audience its own version of part of your message, picked for each customer at send time. See Content variants.	Create a subject line and hero image for each audience.
Personalization tags	Insert a customer value, such as first name or loyalty points balance.	Greet the customer by first name.
Previous interactions	Use what the customer did: cart contents, browsing, and purchases. Built from aggregates.	Show the last product they viewed.
Recommended products	Use output from the recommendation engine, such as bestsellers or similar items.	Show bestsellers from a brand they bought before.
Vouchers	Add a discount code: one assigned to each customer, a unique code from a pool, or one shared code. See Vouchers.	Include a unique code from our SUMMER2026 pool.

In the editor, tags appear as visible Jinja:

{{ customer['first_name'] | default('', true) }}

How Marketing agent builds it

Marketing agent can use any Jinja that's relevant to email and SMS.

Marketing agent tests personalization before you launch, but issues can still happen. For the errors preview can show and how to fix them, see Preview and test and Troubleshoot content issues.

Review and adjust the logic

When personalization uses aggregates, recommendations, segmentations, or other definitions, each one appears as an artifact. Open it to see how it's built, for example which filters the recommendation engine uses.

You can also ask Marketing agent about the logic and change it in chat.

Ask for an explanation:

How does this product block select items?

Ask for suggestions when you're unsure:

What would you recommend for a post-purchase email?

Change the logic:

Change the minimum number of products to 3.

Teach Marketing agent your recipes

Tell Marketing agent how to build recommendations and aggregates for each of your use cases, for example specific catalog filters. It saves them in memory and applies them next time.