The Settings page is divided into tabs (depending on your permissions):
- AI Settings — keys and endpoints for AI features (requires
UWEBCHATVOICE.SETTINGS.VIEW/EDIT) - Security — portal users and roles (requires
UWEBCHATVOICE.SECURITY.VIEW/EDIT)
11.1 AI Settings
Azure AI Speech (transcription)
For automatic transcription of recordings (§7.3) you need an Azure Speech resource.
- In Azure, create a Cognitive Services Speech resource
- Copy the key and region
- Open Settings → AI Settings → Transcription / Speech
- Paste key and region; click Save
The portal stores the key securely. You can remove it with Delete in the same tab.
Azure AI Foundry (summaries)
For AI summaries (§7.4) you need an Azure AI Foundry deployment (e.g. GPT-4o).
- In Azure AI Foundry configure a chat-completion deployment
- Open Settings → AI Settings → AI Summary
- Enter endpoint URL and key
- Click Save
ElevenLabs (Voice Prompt Generator)
For TTS generation (§6.3):
- Create an ElevenLabs account, generate an API key
- Open Settings → AI Settings → ElevenLabs
- Paste key, click Save
Voices can be loaded without the key (public API), but generation requires the key.
11.2 Security tab
Managing portal users
Tenant admins see all contacts of their organisation here from the central customer system. Per user you manage which portal roles they have.
- Open Settings → Security tab
- Find a user (search field with debounce)
- Tick the desired roles on/off
- Changes are synced immediately with the central customer system
Available roles
The portal only shows roles that are specifically created for the Voice Portal:
Universal - uWebChat Voice - Admin(full rights)Universal - uWebChat Voice - SupervisorUniversal - uWebChat Voice - Agent- Custom roles starting with
Universal - uWebChat Voice
Multi-tenant isolation
A tenant admin sees only contacts of their own tenant. The link is made via sipcustomer → account → contacts. Cross-tenant modification is technically impossible.
Audit
Every role assignment or removal is recorded in the Audit Log (§8.3).
Pitfalls
- "Speech key invalid" — wrong region chosen. Check the Azure Portal for the correct pair.
- Summary doesn't work but transcription does — AI Summary endpoint must be configured separately.
- User gets a role but can't do anything — the role is assigned, but logout/login is required to activate the new permissions.