| Unit of deployment | A named coworker hired for one recurring responsibility, defined output, success measure, and human owner. | A private AI assistant for each employee plus a shared teammate in Slack channels. |
|---|
| How work starts | An app event, schedule, request, or handoff from another approved coworker can start the job without someone remembering to prompt it. | Employees can @mention Lindy in Slack, use it in direct messages or channels, and configure scheduled team routines. |
|---|
| Context and learning | Each dedicated team can receive a customized model trained to its use case. The coworker improves from completed tasks and reviewed feedback. | Lindy uses approved team sources, team memory, meeting context, and reusable skills. Users can save work as a skill for the team. |
|---|
| Human control | Configure approvals, prohibited actions, spend budget, named escalation owner, and an auditable timeline per coworker. | Lindy documents named approvers for external actions, central roles and guardrails, source controls, and audit visibility. |
|---|
| Delivery surface | Work can begin and finish across connected business systems, Slack, Teams, email, web, desktop, browser, and MCP clients according to the job. | The enterprise teammate is centered in Slack while acting through connected tools and more than 1,000 advertised integrations, including MCP servers. |
|---|
| Deployment boundary | Managed SaaS and customer-controlled cloud or VPC options are available based on the engagement. | Lindy's public enterprise material describes a managed service with enterprise security controls. Confirm private-cloud, VPC, or self-hosted requirements directly if they are mandatory. |
|---|