Use-case categories

Jobs an agent can run for a business, and the stack layers around those jobs. A category lists a tool only when a verified listing matches it.

How to build a business agent system

• Customer support agents

  Ticket resolution, and a handoff to a person when the agent should stop.

  1 tool

• Sales ops agents

  Lead enrichment, CRM hygiene, and outreach drafts that wait for a person's approval.

  0 tools

• Back office and finance

  Invoice intake, reconciliation, and accounts payable or receivable follow-ups.

  0 tools

• Internal ops and IT helpdesk agents

  Agents for internal operations and IT helpdesk requests.

  0 tools

• Workflow automation with AI steps

  No-code workflows with AI steps, including n8n, Zapier, and Make.

  3 tools

• Multi-agent orchestration and agent SDKs

  Code for orchestration and agent SDKs: LangGraph, CrewAI, the OpenAI Agents SDK, and the Claude Agent SDK.

  4 tools

• Agent runtime and durable execution

  Durable execution, retries, and schedules. Temporal and Inngest are the runtimes listed here.

  2 tools

• Agent memory and context

  Memory and context an agent can read back, including Mem0, Zep, or your own store.

  2 tools

• Evals and observability

  Tracing and evaluation for agent runs, including Langfuse, Braintrust, and Arize.

  3 tools

• Guardrails, approvals, and cost control

  Gateways for guardrails, approvals, and cost control. Portkey is the gateway listed here.

  1 tool

• MCP servers and tool connectivity

  MCP servers and other connections that let an agent call a tool.

  3 tools