---
title: "Version 2.1 Release Notes"
slug: "tea-version-2-release-notes"
tags: ["Release Notes", "Version Release"]
updated: 2026-04-20T17:35:49Z
published: 2026-04-20T17:35:49Z
canonical: "documentation.decisions.com/tea-version-2-release-notes"
stale: true
---

> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.decisions.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Version 2.1 Release Notes

## 2.1 New Features (3/09/26)

| Feature | Description |
| --- | --- |
| Dynamic Tasking | - Expanded Dynamic Tasking capabilities to support the creation of Department-level Task Template Sequences, enabling teams to define and standardize ordered task workflows by department. - Added a **Create Task Template** user action within the Department configuration folder to streamline template setup and governance. |
| Actions on Prompts and Prompt Chains | - Introduced Actions on Prompts to automate follow-up workflows from prompt and chain results, including scheduled reporting, result persistence, and immediate email delivery. - Includes supporting UX updates and configuration controls to manage action creation, frequency, and execution behavior. |
| Canonical Questions | Introduced Canonical Questions to define, manage, and test standardized answers across executions. This feature adds canonical-marking controls, dedicated reporting and test pages, indication strength visibility, and enhancements to execution reports to distinguish canonical results clearly. |
| Execution Report UI/UX Updates | - Improved the Executions Report by removing the clickable link from the Date column and introducing a dedicated Full Response column for accessing execution details. - The response dialog now uses a controlled link for clearer interaction, and the report has been enhanced to include the customer name for better visibility and filtering. |
| Relevance Metrics | Introduced a structured labeling system within prompt results to clearly indicate whether the requested information is present. Each response is automatically evaluated and tagged to distinguish relevant outputs from incomplete or non-responsive results. These visual indicators allow users to quickly assess answer quality, reduce time spent reviewing low-value responses, and confidently identify outputs that meet their original query criteria. |

## Related

- [Text Extraction Agent](/text-extraction-agent.md)