Skip to main content
← All case studies

An AI-powered cross-selling engine built on real sales data

Replacing manually curated cross-selling logic with a data model that recommends genuinely complementary products.

Problem

The existing cross-selling logic on endress.com was defined internally rather than grounded in how customers actually buy, which meant the product recommendations shown alongside a main product were not always genuinely complementary.

Approach

I initiated and led a project to rebuild this logic around real purchasing behavior. Working with SAP sales order data, I drove the development of a calculation model that identifies which products are actually bought together, then used that to power the recommendations shown on the product pages.

Outcome

The result is a cross-selling engine grounded in actual demand patterns rather than internal assumptions, surfacing product pairings that reflect how customers in industrial process automation actually build out their orders.

Working on something similar?

Let's talk about data and AI driven product decisions for your team.

Get in touch