Every seat dollar in your reports uses $30 a seat until you set your real price.
Your Copilot seat price
Who may see the organization's numbers
4 of 12 dials set by your organization
Every answer, narrative, skill and agent draft runs on Aristo-hosted Azure OpenAI.
AI model
The model behind every answer, narrative, skill and agent draft — Aristo-hosted, or your own Azure OpenAI.
Azure OpenAI in Aristo's own Azure subscription. Nothing to set up, included in your plan.
A backup model answers only if your organization has asked for it and the first model fails: the prompt then goes to Anthropic's model through Microsoft Foundry, and Anthropic operates the model and processes the prompt. With the switch off, nothing crosses to the other vendor; a pinned organization never crosses, even with the switch on.
Models on Aristo's Azure: gpt-5, gpt-5-mini, gpt-5-nano.
Your Azure subscription pays for chat; add an embeddings deployment to bill embeddings there too, otherwise they stay on Aristo's Azure as platform-metered usage. Every deployment is checked before saving.
What you would enter: the endpoint from your resource's Keys and Endpoint page, the chat deployment name, an optional embeddings deployment, and the key — which is checked and sealed, never shown again.
This is the illustrative organization — it is never written to, so there is nothing here to set. In your own tenant this control writes the same field the classic settings screen writes, records the same audit action, and every surface that reads the setting moves with it.
Other vendors or private models are arranged through your Aristo contact at your organization's request.