Skip to main content

AI is Part of ADITO Cloud

· 2 min read
Florian Brunner
Florian Brunner
Product Owner @ ADITO Platform

Most of what you hear about AI in enterprise software is vague. Buzzwords, slide decks, promises.

ADITO Cloud ships a self-hosted AI platform. It is running in production today.

This post covers what the platform does, what you can build with it, and where to go next.

What the platform does​

ADITO Cloud includes a self-hosted large language model, accessed through an OpenAI-compatible API. No data leaves the ADITO infrastructure. Everything runs on servers in Germany.

If you have used the OpenAI SDK or REST API before, you already know the interface. The only difference is where the request goes and who controls the data.

Details on endpoints and authentication are in the ADITO-LLM API reference.

What you can build​

The model is not limited to chat. A few things that already work:

  • Structured outputs: ask the model to return JSON in a defined schema and use the result directly in your process logic.
  • CRM-aware prompts: pass record context into the prompt so the model can reason about real data, not generic text.
  • Classification: categorize tickets, emails, or free-text fields without writing rule-based code.
  • Function calling: let the model decide which action to trigger based on user input.
  • Web scraping for context: pull external content into the prompt to give the model more to work with.

Each of these is described in the capabilities overview.

Getting started​

The AI how-to-guide walks you through setup, authentication, and your first request. Start there.

If you prefer to explore on your own, the API is already reachable from any ADITO Cloud system. Point your HTTP client at the endpoint and go.

What comes next​

The roadmap outlines what is planned:

  • A chat UI integrated into the ADITO platform
  • A workflow engine that connects model outputs to business logic
  • Semantic search across CRM data

These are active projects. Expect updates over the coming releases.

One more thing​

AI in enterprise software only matters if it works in your context, with your data, your processes, your constraints.

That is what this platform is built for.

Try it. Break it. Tell us what is missing.

./Flo