Understanding Playbooks

Understanding Playbooks

Playbooks are reusable, multi-step automations in Quickly. They let you define a process once and run it repeatedly, either through conversation, on a schedule, or in response to an event. This guide explains what playbooks are, how to create them, and how to monitor their execution.

What Is a Playbook?

A playbook is a saved sequence of steps that Quickly follows to complete a task. Instead of typing out a complex request every time, you run a playbook by name and Quickly handles the rest.

For example, a "maintenance triage" playbook might:

1. Ask which building the request is for
2. Ask for a description of the issue
3. Determine priority based on the description
4. Create a Jira ticket with the right fields
5. Assign it to the building superintendent
6. Notify the property manager in Slack

Conversational vs. Automated Playbooks

Conversational playbooks run through a back-and-forth dialogue. Quickly asks you questions, you provide answers, and Quickly executes the steps using your input. These are ideal for tasks where the inputs vary each time and require human judgment.

To run a conversational playbook: @Quickly run maintenance triage

Automated playbooks run without user interaction. They are triggered by a schedule or external event, pull data from the trigger source, and execute all steps independently. These are ideal for routine processes with predictable inputs.

The Five Phases

Every playbook follows a five-phase lifecycle:

1. COLLECT - Gather the information needed. In a conversational playbook, this means asking the user. In an automated playbook, this means reading from the trigger event or connected data sources.

2. EXTRACT - Parse the collected information into structured fields. For example, extracting a building address, unit number, and issue type from a free-text description.

3. REVIEW - Present a summary of what is about to happen and ask for confirmation (if the autonomy mode requires it). This is the human-in-the-loop gate.

4. EXECUTE - Perform the actions: create tickets, send messages, update records, fill forms.

5. COMPLETE - Report results and provide links to everything that was created.

Creating a Playbook

1. Go to Playbooks.
2. Click Create Playbook.
3. Give it a name and description.
4. Choose a type: Conversational or Visual Workflow.
5. Define the steps (see below for visual workflow details).
6. Save the playbook.

For conversational playbooks, you define the questions to ask and the actions to take with each answer. The playbook editor provides a step-by-step form for this.

Trigger Phrases

You can assign trigger phrases to a playbook so team members can start it with natural language:

• @Quickly run maintenance triage
• @Quickly new listing
• @Quickly onboard vendor

Set trigger phrases in the playbook settings under Triggers > Phrases.

Scheduling

Automated playbooks can run on a schedule:

1. Open the playbook and go to Triggers > Schedule.
2. Set the frequency: hourly, daily, weekly, or custom cron expression.
3. Choose a timezone.
4. Save.

Scheduled playbooks run automatically at the specified times. Results are logged in the dashboard and optionally posted to a Slack channel.

The Visual Workflow Builder

For complex logic with branching and loops, use the visual workflow builder. It provides a drag-and-drop canvas with 8 node types:

• Trigger - What starts the workflow (message, schedule, webhook, or manual)
• Action - A tool call like "create Jira issue" or "send Slack message"
• Condition - Branch the flow based on data (for example, if priority equals P1, take the urgent path)
• Loop - Repeat a set of steps for each item in a list
• Delay - Wait for a specified duration before continuing
• Human Input - Pause the workflow and ask someone for a decision
• Conversation - Hand off to the AI for a natural language interaction within the workflow
• End - Mark the workflow as complete

Connect nodes by dragging from one node's output to another node's input. The builder validates your workflow in real time and highlights errors before you save.

Running and Monitoring

Running a playbook: Use a trigger phrase in Slack, click Run from the dashboard, or let a schedule or webhook trigger it automatically.

Monitoring execution: Go to Playbooks and click on a playbook to see its run history. Each run shows:

• Start time and duration
• Status (completed, failed, in progress)
• A step-by-step timeline showing what happened at each phase
• Links to any resources created (issues, pages, messages)

Debugging failures: If a run fails, the timeline shows exactly which step failed and why. Common causes include expired integration tokens, insufficient permissions, or invalid input data.

For more on safety controls during playbook execution, see Autonomy Controls.

Building Playbooks in the Dashboard

The guides below walk through the visual editor node by node, which is the part that is hard to picture from a description alone.

How to create a playbook

1. Click Create Playbook and choose whether to create with Quickly Chat or start a blank inactive draft in Studio.
2. Choose the live trigger model the workspace supports, such as manual, message, web form, or scheduled execution.
3. Treat Voice and SMS triggers as telephony-gated: they require live telephony setup before they can receive live traffic.
4. Add the minimum nodes needed to collect information, do the work, and return an output.

How to use trigger nodes

1. Choose the event that starts the workflow.
2. Define any input values the trigger provides.
3. Connect the trigger to the first work node.

How to use conversation and human input nodes

1. Use conversation nodes to collect missing fields from the user.
2. Use human input nodes when the workflow needs approval or judgment.
3. Expect the run to pause and resume after the user responds.

How to use agent and tool nodes

1. Use agent nodes when the step needs reasoning, research, browser work, or multiple tools.
2. Use tool nodes when the step should perform one specific integration action.
3. Feed prior node outputs into the agent or tool so it has the right context.

How to use condition and intent nodes

1. Use condition nodes for explicit rules such as approved vs. rejected or urgent vs. normal.
2. Use intent nodes when the next path depends on what the user meant.
3. Connect each branch to the next node that should run for that outcome.

How to pass data between nodes

1. Use outputs from earlier nodes as inputs to later nodes.
2. Keep field names clear so downstream nodes can reference the right value.
3. When a later node behaves incorrectly, inspect the upstream node result first.

How to review a playbook run

1. Open the run history or Action Logs.
2. Check which node completed, suspended, failed, or skipped.
3. Use the node result to decide whether to edit instructions, input mapping, or branch logic.

Good to Know

• Workflow execution history is visible through Action Logs and workflow run details.
• Plan capacity controls how many playbooks a workspace can keep.
• Members can review and export playbooks, but Manager or Admin access is required to create, edit, activate, delete, or import them.