Connect your AI provider
Power the platform's intelligent features with your own model — choose a provider, supply your key, select a model, and run AI under your own terms and governance.
The platform’s intelligent features — such as the quality and writing checks on requirements — can run on your organization’s own AI model, using a provider and credentials you control rather than an opaque shared arrangement. This guide connects one. For the concepts, see AI Provider.
Connect your model
Bringing your own model is a matter of naming the provider and handing over a key you own.
Before you start In the Identity Portal, go to AI Provider. You'll need administrator access and an API key from the provider you intend to use.
- Choose your provider — OpenAI, Anthropic, or Azure.
- Supply your own API key, and select the model to use.
- Set a custom endpoint URL where your deployment requires one — for instance, routing through a private or regional endpoint.
- Set a maximum tokens per request to bound the size of each call.
Result The platform's AI features run on the model you chose, under your own key. See AI Provider.
Confirm where it’s used
Configuring the provider here sets it for the whole organization.
- Confirm the configured provider backs the platform's AI-assisted capabilities, including requirement quality analysis.
- Because it's set at the organization level, every project's AI features run on the model and terms you've chosen — not on a per-project basis.
Result Consistent, governed AI across every project. See Quality scoring.
Note Bring-your-own is about control: AI calls run under your agreement, your data-handling terms, your spend, and your governance — which is what a regulated organization needs from anything touching its data. If you run a private model deployment, point the platform straight at it through the custom endpoint.
Where to go next
To see the AI-assisted requirement analysis this configuration powers, see Quality scoring.