AI options in VNClagoon+
AI in VNClagoon+ is part of the suite, not an add-on from an outside provider. The same assistant and the same policy apply in every product.
Where the AI runs
Section titled “Where the AI runs”Two choices decide where a request is handled: your organization’s policy, and — on the surfaces that offer it — your own choice.
Your organization decides how AI is provided:
- On your infrastructure — the models run inside your own environment (your data center or a certified European deployment). Nothing leaves your perimeter.
- Approved external model — for a specific use case, your organization may allow an approved external model.
On top of that policy, the Assistant and topic summaries let each person choose, per use, between the shared model and a model on their own computer:
| Option | What it means |
|---|---|
| Server (default) | The request goes to your organization’s model gateway, under the policy above. |
| Local (Ollama) | The request runs on your own computer. Nothing is sent anywhere — not even to your organization’s servers. |
Choose Local for especially sensitive or offline work: Run the assistant on your own computer.
What the AI does
Section titled “What the AI does”- The assistant (Lago) — the pill in the bottom-right corner of every product. Ask a question, or have it summarize the conversation or topic you are looking at, or produce minutes.
- Semantic search — find things by meaning, not only by the words you happened to use. A search for “renewal risk” surfaces the at-risk accounts even when those words do not appear.
- Enrichment — one click to summarize a contact or company, suggest tags, and propose a next action.
Data handling
Section titled “Data handling”With the Server option, requests are handled under your organization’s policy, and a local or on-your-infrastructure deployment never sends a prompt to an external provider. Prompts are sanitized before they reach a model gateway, and the gateway is reachable only inside the deployment.
With the Local (Ollama) option, the prompt never leaves your computer at all: it is sent to a model running on your own machine, and no other person can reach it.