Personalization Agent · Use case
How can a retailer use an AI agent to suggest products to each shopper on its own infrastructure?
A personalisation agent suggests products to each shopper from their browsing and purchase history and your catalogue, within the rules you set for what may be shown. It runs on the retailer's own infrastructure, so shoppers' history is not sent to a public AI service or to an outside recommendation service.
The same page for everyone.
Or shopper data held elsewhere.
- ✗ 01Every shopper sees the same products, or the suggestions come from an outside service that holds your shoppers' data.
- ✗ 02Browsing and purchase history are personal data.
Where it stops for a person
Your team sets the rules and approves each release.The agent suggests and the shopper chooses. Your team sets the rules and approves each version before release, and the agent uses only the data you connect to it.
It suggests products.
Within rules you set.
Step · 01
Purchase and browsing history and the catalogue are connected.
Step · 02
Your team sets the rules: what may be suggested, and to whom.
Step · 03
A shopper visits the site.
Step · 04
The agent suggests products from that shopper's history, within the rules.
Step · 05
Your team reviews the suggestions and adjusts the rules.
Reads from · connectors
The agent connects to Snowflake or a Database for purchase and browsing history, and to your catalogue through your own APIs.
Acts through · MCP servers
It returns suggestions to your site through your own APIs. It changes no price and makes no offer.
What you set up
- Connect the history and the catalogue.
- Write the rules, including what must never be suggested.
- Test the suggestions on past sessions before release.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Does shopper data go to an outside service?
No. The agent and the model run on your own infrastructure.
Can we control what is suggested?
Yes. Your team writes the rules and approves each version before it is released.
Can it change prices or make offers?
No. It suggests products from your catalogue.