General
View the profile's name, touchpoint type, analytics tracking, and fill the Additional Information injected into AI nodes.
The General tab is the first tab of any profile. It defines the profile's identity and a few global settings that apply across all touchpoints.
Profile name
The profile name is for your internal use – it appears in the Profiles list and in the Versions dropdown when assigning a Draft to a specific profile. Use a descriptive name that reflects the environment and touchpoint, for example: Website – English, Retail Kiosk – Warsaw, or Customer Support – Staging.
Touchpoint type
The touchpoint type is set when the profile is created and determines which settings, deployment options, and pipeline features are available. It cannot be changed after creation.
| Touchpoint | Description |
|---|---|
| Web Widget | An embeddable chat interface for any website. Supports text and voice. |
| MetaHuman Kiosk | A full-screen photorealistic avatar on physical kiosk hardware. Supports voice, face detection, and hardware integrations. |
If you need the same virtual being on both a website and a kiosk, create two separate profiles – one per touchpoint type. Both profiles will share the same Pipeline, and Knowledge Base, however they require different Personas due to differences in supported avatar technology.
Analytics
The Analytics field shows the number of active connections on this profile – the real-time count of sessions currently open. This is a read-only indicator, not a configurable setting. For historical analytics, see Analytics view of the Dashboard.
Additional Information
The Additional Information field accepts up to 1,000 characters of free-form text. Whatever you write here is automatically injected into the system prompt of every LLM Response and Tool Agent node in the pipeline.
Use this field for context that should be universally available to the AI regardless of which flow is active. Good candidates include:
-
Where this deployment physically or digitally sits (e.g. "This kiosk is installed in the main lobby of Building 3, next to the visitor check-in desk.")
What this touchpoint is, if it behaves differently from others (e.g. "This is the kiosk version of the assistant — users speak aloud rather than type, so keep responses short and easy to follow by ear.")
A time- or event-bound context (e.g. "This assistant is deployed for the duration of the 2026 Product Expo, at the entrance to Hall B.")
The current deployment context (e.g. "This assistant is deployed on the customer support page of example.com. Users arriving here have already purchased a product.")
Example:
You are deployed as a kiosk in the main entrance of Acme Corp's manufacturing plant in Detroit. Visitors speak to you aloud rather than type, so keep responses short and easy to follow by ear. This kiosk is active during the plant's public Open House week — most visitors are prospective clients touring the facility, not existing customers.The Additional Information content is passed to the LLM on every turn. Keep it concise and high-signal – long paragraphs of general background will consume token budget without meaningfully improving responses. Aim for 2–4 tightly written sentences covering what the AI must always know.