Search playbooks in chat

Note
Preview feature: Playbooks are currently in preview. Functionality may change in future releases.

When playbooks are visible to agents, the Assistant automatically searches and references them when answering questions—even if you don’t mention playbooks explicitly.

You can also search playbooks manually in chat using the context menu. Type @ in the chat input to open the context menu and select playbooks to reference.

How semantic search works

Playbooks are indexed in a semantic vector database. The system breaks playbook content into chunks and creates vector embeddings that capture meaning and context. When you ask a question, the Assistant searches these embeddings to find the most relevant sections—even if your exact words don’t match the playbook text.

Agents retrieve specific parts of playbooks as they work, not entire documents. This means agents can find and use relevant troubleshooting steps, configuration details, or procedures from different playbooks in a single response.

Visible to agents: Only playbooks with Visible to agents enabled are indexed and searchable. Enable for shared knowledge; disable for drafts or commands.

When playbooks are searched

The Assistant automatically searches playbooks when your question requires:

Domain-specific knowledge about your systems or processes
System architecture or design information
Troubleshooting guidance
Procedural information

Use with Investigations

Investigations automatically search and follow playbooks when enabled. When you launch an investigation, specialist agents retrieve relevant playbook sections as they work, following your documented procedures and troubleshooting steps.

This automation means Investigations can:

Follow your established playbook without manual guidance
Apply consistent procedures across different incidents
Retrieve alert-specific response steps automatically
Combine knowledge from multiple playbooks in a single investigation

To enable playbook search in investigations, ensure playbooks have Visible to agents enabled. Investigations respect the same scope settings as chat.

Optimize for search

Structure playbooks to help agents find the right information quickly.

Descriptive titles: Use clear, specific titles like “Investigate Database Connection Failures” instead of “DB Stuff”.

Include keywords: Mention specific service names, metric names (e.g., CoreDNSErrorsHigh), error codes, and process names.

Clear structure: Use headings and step-by-step instructions.

Organize by alert: Structure playbooks with sections per alert name. This helps agents find the exact troubleshooting steps for specific alerts. For example:

# Infrastructure Troubleshooting Playbook

## CoreDNSErrorsHigh

When this alert fires, check:

- DNS query latency in the CoreDNS dashboard
- Pod restart count in the last 15 minutes
- Network policy changes that might block DNS traffic

## DatabaseConnectionPoolExhausted

When this alert fires, check:

- Active connection count vs pool size
- Long-running queries blocking connections
- Application connection leak patterns

Each section is indexed separately, so agents can retrieve only the relevant alert response steps.

Scope

Search respects playbook scope: tenant-scoped playbooks are discoverable by all users; user-scoped playbooks only by the creator.

Troubleshooting

Not found: Ensure “Visible to agents” is enabled and wait for indexing.
Irrelevant results: Use more specific terms, service names, or error messages.