E-commerce & info business

First interest-graph network

Almost 10,000 users connected through their interests with AI.

Almost 10,000 users discovered content and people with similar interests through the network without having to search for them specifically. The engine was subsequently handed on.

users connected through their interests
~10,000
users connected through their interests
Abstract data streams

Overview

01 · Problem

Relevant content and contacts were hard to discover without a clear search query.

02 · Approach

An interest model matched content and users with one another.

03 · Result

The network reached almost 10,000 users; matching content and contacts became easier to discover.

Discuss a similar bottleneck →

Before / after

What changed in the business.

DiscoveryBeforeOnly with a clear search queryAfterSuggestions by thematic proximity
ContactsBeforeBy chanceAfterPeople with similar interests
ReachBeforeNew productAfterAlmost 10,000 users at hand-over

E-commerce & info business

Linking semantic content and contacts

Starting point

For a new digital product, people were meant to find relevant content and contacts without knowing exactly what they were looking for.

How do interest-based content recommendations work?

During the project, an early AI-based interest-graph network and an anonymously usable interest space were built. The platform recognises thematic proximity between visited content and suggests further websites as well as contacts with similar interests.

Before the engine was handed on, the network reached almost 10,000 users.

At a glance

Levers and services.

This case is not clearly assigned to any of the five levers; what matters is the type of solution and the industry.

Services
StrategyImplementation
Type of solution
AI & company knowledge

Result

Result and assessment.

users connected through their interests
~10,000
users connected through their interests

Almost 10,000 users discovered content and people with similar interests through the network without having to search for them specifically. The engine was subsequently handed on.

What entrepreneurs can take from this

Recommendations also work without a purchase history when the system understands what the content is about.

A mechanism, not a promise of results: the effect depends on your starting process.

What this means for you

Building blocks from this case.

This project gave rise to use cases that we now deploy in other companies, adapted to their processes.

Frequently asked questions

Questions about this case.

Your question is not here? In the initial conversation we take 30 minutes for your specific topic.

It was handed on. By then the network had reached almost 10,000 users.

Almost 10,000 users discovered content and people with similar interests through the network without having to search for them specifically. The engine was subsequently handed on.

The principle yes, the implementation never one to one. Recommendations also work without a purchase history when the system understands what the content is about. In the initial conversation we check which part of it has the greatest leverage for you.

More case studies

All case studies →

Should customers find what they are not searching for?

In the free initial conversation we look at your process, show a comparable project and assess where the greatest lever lies.

Free · 20 minutes · online

Your next step

Discuss recommendation logic →Start with the Digital check-up →