Zum Inhalt springen

Glossary

Dieser Inhalt ist noch nicht in deiner Sprache verfügbar.

The AI counterpart of one human expert. A SynTwin captures the owner’s knowledge, communication style, and decision-making.

The human expert behind a SynTwin. The relationship is 1:1 — one human, one SynTwin.

Puts a SynTwin into use by setting context and goals. Others access a SynTwin through a deployment. Think of it like the same expert switching between “mentor mode” when coaching juniors and “marketing mode” when speaking at events — same person, different context. A deployment is such a context.

A deployment can provide initial context and a goal to shape how the SynTwin behaves. When a session achieves its goal, that information surfaces in the dashboard.

A conversation between a human and a SynTwin. Ends when the visitor clicks the end session button or after a timeout. Each session is evaluated against its deployment’s goal.

A known individual connected to sessions and sources. Visitor sessions create or attach to a person record, which may start without a name or email. The SynTwin can fill in details learned during conversation. The owner can manage people in the app.

The SynTwin retains memory of past conversations with and about each person and connects the dots across sessions.

Material that feeds a SynTwin’s knowledge.

  • Global sources are owner-level sources that deployments can opt into.
  • Person-specific sources apply to one person.
  • Style-bearing sources help extract communication style.

Supported uploads include PDFs, docs, slides, Markdown, text, and CSV files.

The umbrella term for extracted patterns that shape how a SynTwin thinks and communicates. Rules come in two kinds:

  • Knowledge rules — higher-level factual summaries that are always top of mind for the SynTwin. Think of these as things the expert just knows without looking anything up.
  • Style rules — patterns of conversation style extracted from uploaded conversations and refined through 1:1 sessions.

In the UI, rules are filtered by kind using the “Knowledge” and “Style” tabs.

The channel a deployment supports — chat, audio, or video. A deployment can enable one or more modes depending on the desired experience.

An objective set on a deployment that defines what a successful session looks like. When a session achieves its goal, that outcome is tracked and surfaced in the dashboard.

Deployment-specific behavior guidance. Instructions shape how the SynTwin should pursue the goal in that context.

Information provided to the SynTwin at the start of a deployment session — for example, details about a lead or a customer’s history. Shapes how the SynTwin approaches the conversation.

The screen visitors see when a deployment supports multiple interaction modes. Lets them choose before starting a session.

A label on a source. Tags are auto-extracted during processing but can also be created and managed manually by the owner.

A recording of the owner’s voice used to clone how the SynTwin sounds. Multiple voice samples can be uploaded, and a specific sample is selected per language the SynTwin supports. This allows the SynTwin to not just talk like the owner, but sound like them too.

A photo, recorded video, uploaded video, or external replica used for video deployments.

The post-session improvement loop. SynTwin can create or update rules, raise questions, and suggest instruction changes after visitor sessions.

An action card the SynTwin can show during a deployment session when configured conditions are met.

A deployment link that carries known visitor context such as name, company, email, and a short context note.

A deployment link for an existing person. It lets the next session resume with that person’s memory.