General enhancements

You can override the AI provider and model used by Generative AI activity directly in the process. This lets you select the provider or model best suited to a specific activity based on cost, speed, or accuracy without being limited to the default set at the Integration or System settings level and without republishing the solution when the choice changes.

The same provider/model override pattern available for process activities now extends to forms. You can override the provider and model used by a Generative AI chat control, Copilot control, or form action via the same dropdown or inline-value approach. As a result, a chat or Copilot experience on a form can use the provider and model best suited to it, independent of the Integration's or System settings' default.

You can now override the model on a template's Generative AI activity using skins.

Refer to Create a skin in the TotalAgility Advanced Studio Help and Create skins in TotalAgility Studio Help.

The Generative AI Chat control now exposes two additional properties, Prompt and Conversation, giving full access to the conversation flow when "Persist conversation" is enabled.

TotalAgility provides centralized certificate management for RESTful service activities. You can upload and store certificates in a central repository and reference them by name in RESTful service activities that use JSON Web Signature Token custom headers.

When exporting processes, the export package includes the names of the referenced certificates, not the certificate files themselves. Upload the required certificates to the target environment before importing the process.

Refer to Certificates in the TotalAgility Advanced Studio Help.

When creating a global or local data model from a sample JSON file, designers can automatically populate default values for Text, Number, Date, and Boolean fields by selecting the Populate defaults from sample JSON option. This saves time by eliminating the need to enter default values for each field manually.

Refer to:

• Data models
• Create a global data model using JSON
• Create a local data model using JSON

The Data list activity configuration includes an explicit All option for Delete rows, Update field values, and Get rows operations. This makes it clear when an operation will affect all rows in a data list, preventing accidental bulk operations that may occur when search criteria are unintentionally left empty.

Any existing Data list activity configurations with empty search criteria are automatically converted to use the All option, preserving existing functionality.

Refer to Configure a Data list activity in the TotalAgility Advanced Studio Help.

To enhance security, File upload control settings are no longer transmitted as HTTP request parameters from the client to the server. Because the upload settings are sourced from a trusted server-side repository rather than the client, they cannot be modified or tampered with by users. This approach strengthens enforcement of upload policies and improves the overall security of the file upload process.

When configuring a Data activity or form action, a list of database schemas (based on the user in the database connection string) is available to select a database schema to use for retrieving database objects.

The new Clear on archive property available for Strings, Nullable Strings, Data Objects, and Dynamic Complex variables helps reduce database storage requirements by automatically clearing variable values when jobs are archived. This feature is particularly beneficial for cloud deployments where database size can directly affect costs and system performance.

When configuring a Separation profile, the new Advanced tab is available for configuring advanced bar code settings for Code 39 and Data Matrix bar code types using a JSON string.

Generative AI integration configuration is simplified for administrators. When configuring Generative AI integration and selecting Azure OpenAI as the type, you no longer need to specify an API version in the URL; it is ignored if specified. Across all Generative AI providers, the API version and URLs are now applied automatically based on the configuration.

When importing a package with a deployment type, you can overwrite the transformation settings in the deployment configuration on the target system. By default, those settings are not overwritten, and only additional transformation settings are added to the artifact list.

The maximum length of a case reference is increased from 40 to 50 characters.

When configuring the Tungsten AI provider to use OAuth server authentication, you can define custom headers to enable access through proxy servers that require additional header information. These headers can be used to pass authentication credentials, client identifiers, or other values required by the proxy, ensuring secure and seamless connectivity to the Tungsten AI provider.

Refer to Configure Tungsten AI provider.

Two new SDK functions, "Stream GetPageImageByIndex" and "String GetPageImageByIndexAsBase64" have been added that return the image of a page within a multipage document, using the page index instead of the page ID.