Configure collaboration, automation, marketing, contact management, and call operations.
AI Agents
AI Agents is where you create, configure, and monitor the AI workforce for a workspace.
Main AI Agent Areas
Agents
Use AI Agents > Agents to create and manage agents. Start with focused agents that have one clear job, then add skills and integrations only when the job requires them.
Memory
Use AI Agents > Memory to maintain shared context. The Memory area includes setup, business details, brand and voice, strategy, guidelines, knowledge, and Memory Palace entries.
Use it for:
- business identity, services, locations, and audience
- brand voice and operating guidelines
- goals, competitive notes, and strategic context
- uploaded or linked reference material in
Knowledge - durable notes in
Memory Palace
Skills
Use AI Agents > Skills to manage capabilities agents can use. Skills should match real work the agent is expected to do.
Voice
Use AI Agents > Voice for voice-related operations when available. The Voice area includes calls, numbers, A2P setup, templates, playground testing, and settings.
Agent History
Use Agent History to inspect past agent runs and understand what happened during an automation or conversation.
What To Configure On Every Agent
Every production agent should have:
- a clear name and job description
- a specific goal for the work it owns
- relevant shared context from
Memory - only the skills and integrations it needs
- approval rules for sensitive actions
- a channel where teammates can review outputs
Shared Context vs Agent-Specific Setup
Use AI Agents > Memory for information every agent should inherit.
Examples:
- company services
- locations and service areas
- brand voice
- pricing guardrails
- approved offers
- review-response rules
- compliance notes
Use the individual agent configuration for what only that agent needs.
Examples:
- campaign-specific goals
- a narrow role description
- selected skills
- connected tools
- channel assignment
Safe Rollout Pattern
- configure shared memory
- create one focused agent
- connect only required integrations
- test in a private or internal channel
- require approval for write actions
- review Agent History after the first tasks
- expand access only after the team trusts the behavior
When To Use Voice And Heartbeats
Use Voice when the agent needs call-related setup or testing. Use Proactive Heartbeats when the agent should check for changes on a schedule and alert the team.