SUPPORT.TWILIO.COM END OF LIFE NOTICE: This site, support.twilio.com, is scheduled to go End of Life on February 27, 2024. All Twilio Support content has been migrated to help.twilio.com, where you can continue to find helpful Support articles, API docs, and Twilio blog content, and escalate your issues to our Support team. We encourage you to update your bookmarks and begin using the new site today for all your Twilio Support needs.

How do I create a custom language operator?

Objective

This article walks you through building your own language operator when no existing operator covers your use case, writing its instructions, defining any parameters and output format, adding training examples, and choosing whether it can use extra context.

 

Product

Conversation Intelligence

 

Environment

Twilio Console

 

Procedure

  1. In the Twilio Console, select Orchestration in the left sidebar.
  2. Select Conversation Intelligence, then Language operators.
  3. Select Create operator.
  4. On Create instructions:
  5. Enter the operator name. This is required.
  6. Optionally enter an Operator description.
  7. Write the prompt describing the task this operator should perform with your conversations. This is required, be specific and include the key requirements for how it should behave, because more detail produces better results.
  8. Optionally use Add parameters (optional) to create custom variables you can reference in your prompt. Their values are set per rule inside your Intelligence configurations.
  9. Choose the output format: Text for a plain-text result such as a summary or response, or JSON for results a third-party application will ingest programmatically.
  10. If you chose JSON, provide the output schema. A link to the schema guidelines is shown next to the field.
  11. Under Add training examples, add input-and-expected-output pairs to teach the operator how to analyze transcripts. This makes results more accurate and tailored.
  12. On Allow context capabilities, choose whether this operator may use extra context:
  13. Make Enterprise Knowledge capable lets it recall organization-specific information such as policies, FAQs, and internal documentation.
  14. Make Conversation Memory capable lets it use conversational observations and customer profile traits.
  15. On Summary, review the operator's details and finish. A confirmation appears that the operator was created.

 

Additional Information

Allowing a context capability does not switch it on. You choose exactly how context-aware each rule is when you create or edit the rule in an Intelligence configuration, the operator simply becomes eligible.

Every field you set here can be changed later from the operator's details page.

If the JSON schema you enter is not valid JSON, you are asked to correct it before continuing, review the schema guidelines if you need the expected format.

Once created, the operator appears in the Language operator library and can be added to any rule.

 

This content was generated by AI and reviewed, edited, and verified by a human.

Have more questions? Submit a request
Powered by Zendesk