Settings
Browser credentials for API access, connector secrets, and AML screening notification settings.
Investigation Controls
Configure LLM prompts and enable or disable controls for your investigation workflow.
| Num | Name | Type | Enabled/ Report | Actions |
|---|
Configure who receives AML screening hit reports. Reports are sent automatically when a screening job finds matches.
This is how screening hit reports will appear. Template customisation coming soon.
How long an action plan approval token stays valid before it expires. Each tenant sets this independently.
Default: 300 (5 min). Range: 60 – 86 400. Staff who take longer to review plans need a higher value.
The AI provider and model this tenant uses for NL commands, investigation controls, and RAG retrieval.
Required for Azure OpenAI and self-hosted models.
Only used when provider is Azure OpenAI.
Used to convert document text and queries into vectors. Changing this after documents are ingested requires a full re-embed.
Must match the model's output dimension. text-embedding-3-small = 1536.
Browser origins permitted to call this tenant's API. One origin per line.
Generate the key on Enterprise Memory → Tenant Register, then paste the Microsoft tenant ID and secret here. AML does not create this key.
Sent to Enterprise Memory as X-MS-Tenant-ID.
Leave blank when saving to keep the existing key.
Optional. Public https base URL for Enterprise Memory (the host other servers will call). Overrides EM_SERVER_URL for this project; leave blank to use the server environment. In development mode, localhost, 127.0.0.1, and host.docker.internal with http or https are accepted. Do not use Kubernetes service names or *.svc.cluster.local in production — AML must keep working if moved to another server. Test connection calls the same EM Search and Sources APIs as Test Query.
| Label | Type | Created | Actions |
|---|