How it works

Transparent architecture. No black box. You understand every layer.

Your Systems CRM, SAP, Email, ERP Your Data Customers, HR, Finance SECURITY LAYER ✓ Data Masking ✓ Access Control ✓ Audit Trail ✓ Provider Route OpenAI Claude Local (Ollama)

LAYER 01

Your Systems

Which systems can be connected?

  • CRM
  • ERP (SAP)
  • Email (Microsoft 365)
  • Ticketing (Xurrent, ServiceNow, Jira)
  • HR Systems
  • Financial Software

Integration via REST API, webhooks, or direct database access. All connections encrypted (TLS 1.3).

CRM
ERP (SAP)
Email (Microsoft 365)
Ticketing (Xurrent, ServiceNow, Jira)
HR Systems
Financial Software

security-layer.config

Data Masking

Personal data (names, emails, account details) are automatically masked before being sent to AI APIs. Only the Security Layer knows the real data.

Access Control

You define exactly: which agent can access which systems. No agent sees more than it needs. Principle: Least Privilege.

Audit Trail

Every action of every agent is logged. Who, when, what, with which data. Exportable for compliance audits (GDPR, SOC2).

Intelligent Routing

Sensitive data → local LLM. Non-critical tasks → Cloud API. You set the rules, the Security Layer enforces them.

LAYER 02

Security Layer

Detailed control over every data flow.

LAYER 03

LLM Provider

Which providers are supported?

Provider Type Strength Best for
OpenAI (GPT-4o) Cloud Broad knowledge, code Non-critical tasks, code generation
Anthropic (Claude) Cloud Analysis, reasoning Complex document analysis
Ollama (Llama 3) Local Privacy, control Sensitive data, on-premise
Custom Models Local/Cloud Specialized Industry-specific tasks

Ready for an initial conversation?

In a 30-minute initial call we listen, assess your situation and give you an honest take on whether an AI agent makes sense here — or not.