| Agent | An AI-powered voice assistant built on Atoms. Agents handle phone calls or web conversations based on your configuration. |
| Single Prompt | An agent type using one comprehensive prompt for flexible, free-flowing conversations. |
| Conversational Flow | An agent type using a visual workflow for structured, goal-oriented conversations with defined paths. |
| Node | A single step in a Conversational Flow workflow. Types include Prompt, Transfer Call, API Call, End Call, Pre-Call API, and Post-Call API. |
| Branch | A connection between nodes in a Conversational Flow that determines the next step based on conditions. |
| Cold Transfer | A call transfer where the receiving party is connected immediately without any context or briefing. |
| Warm Transfer | A call transfer where the AI provides context to the receiving party before connecting the caller. |
| Whisper Message | A message spoken only to the transfer recipient (not heard by the caller) during warm transfer. |
| Knowledge Base | A repository of documents, FAQs, and information that agents can reference during conversations. |
| Webhook | An HTTP callback that sends data to your systems when specific events occur (call start, call end, etc.). |
| PII Redaction | Automatic removal or masking of personally identifiable information from transcripts and logs. |
| Campaign | An outbound calling program that contacts a list of phone numbers using your agent. |
| Audience | A contact list (phone numbers and optional data) used for outbound campaigns. |
| Disposition | The outcome category of a call (e.g., successful, unsuccessful, transferred). |
| Post-Call Metrics | Analytics and data extracted from completed calls (satisfaction scores, issue types, etc.). |
| STT (Speech-to-Text) | Technology that converts spoken audio into text for processing. |
| TTS (Text-to-Speech) | Technology that converts text responses into spoken audio. |
| LLM (Large Language Model) | The AI model that understands intent and generates responses. |
| SLM (Small Language Model) | Atoms’ optimized models that deliver fast, efficient voice AI with lower latency and cost. |
| Latency | The delay between when a user finishes speaking and when the agent starts responding. |
| Turn Detection | The system that determines when a speaker has finished their turn and it’s time to respond. |
| Interruption | When a caller speaks while the agent is talking, requiring the agent to stop and listen. |
