# March 24

## Platform Improvements

### **New Globals and Accounts tab on the Build page**

Accessing and configuring **Globals** and **Accounts** is easier now. New dedicated tabs are available on the **Build** page, allowing you to create and manage global variables and accounts directly from the area where pipelines are built.

Previously, Globals and Accounts could only be configured through the Settings page. With this update, you no longer need to navigate away from the Build context, making the setup process faster, more intuitive, and better integrated into your workflow.

<a href="https://app.gitbook.com/s/jvO5S91EQURCEhbZOuuZ/development-cycle/build-overview" class="button primary">Learn more</a>

<figure><img src="https://4091659634-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FE9x0KHkAmxzJGlYPs9aE%2Fuploads%2FwfbWjOuuA4zRt8NwEdy5%2Fglobals-and-accounts.gif?alt=media&#x26;token=f3ecd2e2-e02b-4a41-9111-85dbd31ec6cf" alt=""><figcaption></figcaption></figure>

## Connectors

### **New Double Braces function: ISABSENT**

The new `ISABSENT()` function simplifies how you handle missing data in Double Braces expressions. It automatically normalizes null, empty, or non-existent values into a single `null`, eliminating complex workarounds and making your pipelines cleaner, more reliable, and easier to maintain.

<a href="https://app.gitbook.com/s/EKM2LD3uNAckQgy1OUyZ/double-braces/double-braces-functions/conditional-functions" class="button primary">See how the function works</a>

### **Blob Storage connector now supports Dynamic Accounts**

The Blob Storage (Azure) connector now supports Dynamic Accounts, allowing credentials to be defined and updated at runtime instead of being statically configured. This makes pipelines more flexible, simplifies credential management, and enables dynamic authentication scenarios, such as temporary tokens or runtime-provided credentials, with greater ease and security.

<a href="https://app.gitbook.com/s/EKM2LD3uNAckQgy1OUyZ/connectors/file-storage/azure-blob-storage" class="button primary">Read the documentation</a>

## Documentation

Ever wondered how to make sure your AI is actually doing what you expect? We added two new guides that walk you through building an automated quality gate for AI-generated tasks and then making it smart enough for production:

* [**Automate AI task validation with a quality gate**](https://app.gitbook.com/s/aD6wuPRxnEQEsYpePq36/quickstarts/automate-ai-task-validation-with-a-quality-gate): Build an Evaluator Agent, test it against mocked data, and see where a strict prompt breaks down.
* [**Optimize the AI auditor with semantic intelligence**](https://app.gitbook.com/s/aD6wuPRxnEQEsYpePq36/quickstarts/optimize-the-ai-auditor-with-semantic-intelligence): Fix the prompt using Versioning, make the evaluator semantic, and confirm the improvement with the same test Dataset.

## **Bug fixes**

* **Agent Component fixes:** Fixed an issue where Datasets and Evaluations were not shared across minor versions of the same pipeline, ensuring consistent access to resources throughout the pipeline lifecycle.
* **Output Schema from MCP Tools:** Fixed a bug where the Output Schema was not displayed in the tools/list operation for the MCP Server, ensuring tool schemas are now correctly displayed.