Try AI

Try AI is a simple-to-use tool enabling back office Capture users to see precisely how AI can caption and keyword still images from their portfolio. It is now available on all Capture back office systems.

Try it with 100 credits for free!

If you are using the Capture Back Office and switch to New Generation Capture Beta One, the 'Try AI' button will appear at the top of the Media Browse & search screen. This short article describes how to use this new feature.

The tool is designed to deliver a single, easy-to-use interface to multiple AI engines*. This means less time and complexity for Capture users to gain a really solid understanding of the value AI can deliver for the type of content you manage in your Capture system.

Users can purchase credits in order to use the AI engine. A credit is consumed every time an image is processed by an AI engine to produce a caption and keywords. These results can then be stored in Capture metadata fields as required.

For 1,000 images to be processed requires 1,000 credits. These can be used across any of the AI engines that are available in the system.

*Please note that at the time of launch a single AI engine is included from Microsoft Azure AI Services, with more to follow in due course.

How is data retained and what controls are available?

Try AI uses Microsoft Azure AI services by default. The images submitted to this Image Analysis service are low resolution (preview size) and are processed in real time. The input images and results are not retained or stored in the service after processing.

To learn more about privacy and security commitments, see the Microsoft Trust Center.

For additional security best practices and information, visit Azure AI services security baseline.

The Try AI Screen

To see the 'Try AI' screen, switch to New Generation Capture Beta One using the toggle button at the top of the back office screen. Go to Media > Browse & search, select some assets and click on the 'Try AI' button which should appear top left next to the 'Download' button.

The initial view (see the screen shot below) displays the Try AI heading top left, the Credit Balance top right. In a bar below these there is a button labelled 'Generate', a dropdown list with the default value 'Merge Existing And AI Data', and two more buttons: 'Apply To Selected' and 'Save All Manual Edits'.

To the right of these buttons you can see the number of images selected and the number of credits that will be deducted after generating AI content for these images. Below this bar there are rows for each image with 3 columns showing image thumbnails, existing content  and AI generated content. Each image thumbnail has a checkbox on the left to show if it is selected.

The 'Existing Content' column shows two fields which can be configured as required. In the screen shot below these are the Description and imported Keywords fields. You can edit text in these fields. The 'AI Generated Content' column shows the Description and AI-generated keywords for each thumbnail.

There is also a link below the Credit Balance which you can use to purchase more credits.

Before using Try AI for the first time, you can configure the AI recommendations settings. Click on the Settings icon shown below to view the configuration dialog box. This icon is in the 'AI Generated Content' header area.

You can then select which two fields to view on the left, currently showing Description and Keywords.  Alternatively these might be two fields created to hold AI-generated content for comparison with existing content later. You can copy the AI-generated description and keywords to the metadata fields, as we will see below.

You can also select the AI engine at this point if a choice is available, see the screen shot below.

Image Selection

Select images for captioning and keywording with AI by clicking on the image thumbnails or checkboxes. You can also tick the checkbox at the top to select all images (see screen shot below). After selecting images you will see the number selected and the number of credits that will be deducted after content is generated. The 'Generate' button will also be activated, as shown below.

Generating Content

When the images have been selected click on the 'Generate' button to display AI generated content. This will be shown in the 'AI Generated Content' column. If any assets are not images and AI content cannot be generated, an error message will be shown in the Keywords box for that asset.

After AI content is generated, you can select from the dropdown the option to merge existing data with AI data or to replace existing data with the AI content.

Note that if you return to the Try AI screen with assets that have been selected previously, the system will remember if the assets had any AI content generated earlier. In that case, you will not be able to use the AI generator again with those assets, so there's no risk of accidentally wasting credits that way.

Saving the AI-Generated Content

From the dropdown, you can select the 'Merge Existing And AI Data' option to merge the data, or if you want to replace existing content, select 'Replace Existing With AI Data' from the dropdown. For either option, click on 'Apply To Selected' (see below) to make the changes.

A warning will be displayed as shown below. If you click on 'Update' to confirm, the existing data will be overwritten or merged with the AI generated content, depending on your selection.

Merged data will combine the AI generated content with the existing data, as shown below.

At this stage, you can edit the fields on the left if required. If you do this, the 'Save All Manual Edits' button becomes active (see below). Click on this to save your changes. This Save button is only needed if you edit the fields manually. When the AI content is copied across, it is saved automatically.