Intel / article

Procedures

Source: elevenlabs.io
Author: ElevenLabs editorial sitemap
Date: Published Oct 12, 2026

Context

Procedures give ElevenLabs' conversational agents task-specific instructions that load when triggered. A structured Procedure specifies a fixed sequence of steps, suited to a consistent intake process. A free-form Procedure supplies and boundaries while allowing the to adapt its actions during a conversation. Teams can import an existing operating procedure as a document and review the draft that ElevenAgents produces. The announcement describes this as a way to manage individual tasks without putting all their instructions into one broad . Procedures are in Alpha.

Key takeaways · AI-distilled

Structured Procedures run a fixed sequence of steps the same way every time, built from natural language, tool calls and instruction chips. ElevenLabs positions them for consistency-critical tasks such as claims intake and payment collection.
Free-form Procedures supply task context, steps and , and the agent interprets them and adapts to the conversation. ElevenLabs suggests them for tasks that shift as they go, such as technical troubleshooting and billing questions.
Teams can import existing SOPs as docs, PDFs or txt files and ElevenAgents drafts a Procedure for a person to confirm or refine. Procedures sit alongside the system prompt, workflows and guardrails rather than replacing them; the feature is in alpha.

Terms in this piece · Glossary

context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
system prompt — The standing instructions a model receives before any user input — defining its role, rules, tools, and tone for the whole conversation.
guardrails — The checks around a model that block bad inputs and outputs — filters, validators, and permission rules the model itself can't override.

Read the source elevenlabs.io