Agents that know when to ask

Use Cases

• Data Analysis Agent
• Support Agent
• CRM Agent
• Meeting Prep Agent
• Call Analysis Agent

Data Analysis Agent

Get instant answers, directly from Slack

Support Agent

Automatically triage issues and spot patterns.

CRM Agent

Manage your CRM without clicking into your CRM.

Meeting Prep Agent

Triage, summarize and resolve requests

Call Analysis Agent

Understand what's working through natural conversation.

When to loop in a human

You might have heard the term “human in the loop” before to describe these kinds of hybrid human-AI agentic processes. Humans intervene at critical moments to make the most impactful decisions, while agents do the rest of the work behind the scenes.

An agent can alert you and ask for your approval when it’s:

• About to do something customer-facing: an agent drafts a reply to a customer ticket in Zendesk, but waits for your sign-off before actually sending it.
• About to do something irreversible: an agent stages a batch of Salesforce records to be deleted, or payments to send, and waits for your approval before taking action.

How to bring a human into the loop

For any app that you have connected to an agent, you can control the approval settings: allow every tool call without asking, ask before every tool call, ask before write/delete tool calls, or, you can get even more granular with custom settings.

For example, you could allow an agent to run most Gmail tool calls without requiring approval, but require human approval before the agent actually sends an email.

You can see the requests pending approval for a particular agent in a "Pending Approval" section just above your recent chats with that agent.

Human in the loop features make it possible for you to give your agents the autonomy they need to get routine work done, while also allowing you to stay as the decision-maker on the moments that matter.

Automate any workflow

with AI

No coding required.