Project showcase · What we built

Email automation with machine learning

5,000+ emails processed automatically every day with ML routing

Starting point and solution

The challenge

Over 5,000 emails a day tied up the team in sorting work. Little time was left for complex customer requests and advice.

Our solution

Email processing with machine learning: the system classifies incoming emails automatically and routes them directly into the existing systems.

Staff focus on complex customer advice
Team relieved of repetitive sorting work
Faster response times for complex requests
Misrouting rate reduced from 8% to 5%

The process, step by step

The diagram replays a typical run, from intake through processing to handoff into the target systems. It shows the structure of the delivered system. Nothing runs live on this page: it is a diagram, not a view of live operation.

Core features

The main components of the solution

Automatic email routing

ML-based classification at 95% accuracy. Automatic distribution into 5+ categories with fallback mechanisms.

AI-assisted extraction

An LLM automatically extracts relevant data from emails for direct downstream processing.

Direct SAP connection

Processes are created directly in the ERP system without manual entry.

Live monitoring

Real-time visualization of all email flows and metrics.

Scalable architecture

In production with 5,000+ emails a day, with headroom for peak loads.

GDPR compliant

Full encryption and audit trails.

Technology stack

The technologies and building blocks used in this project

GPT-4 TurboCustom ML ModelLangChainSAP IntegrationApache KafkaKubernetesWebSocketPrometheus

A comparable process in your organization?

We check in a short readiness assessment whether and how your use case can be implemented, with a clearly bounded scope.

Schedule a conversation See more projects