AI Integrations in Campaign QA Checklist

Updated 

Eliminate the risk of off-brand advertising with an automated QA Checklist. 

The integration of AI Checks in QA Checklist provides wider governance for your creatives by verifying essential media asset details covered by AI engines for human presence, face presence, language and much more. In this article, we will guide you through the AI capabilites supported in Campaign QA checklist to ensure brand compliance and minimize the risk of off-brand advertising.

Enablement Note: The use of this capability requires that Sprinklr Advertising be enabled in your environment. To learn more about it, please work with your Success Manager. 

List of AI Checks in QA Checklist

AI Check Fields

Description

Language detected in asset & copy

Detects language of the text used in the asset

Aspect ratio of Media

Verify aspect ratio based on channel guidelines

Human present in Media

Validates presence or absence of human in the media asset

Prominent color in media

Detects the top 5 colors present in the asset

Word count bucket of the Headline in the Media

Ensure media is not text-heavy and conveys the intended message effectively

Face present within first 5 seconds of video

Verify if face is present within first 5 seconds of the video to check for emotional appeal aspect

Sound in video

Check for any issues in the sound of the video asset

Subtitles in video

Accessibility check to ensure subtitles are included when needed

Video length bucket (seconds)

Categorizes videos according to their duration into different groups

Audio language in video

Detects the audio language used in the asset

Subtitle language in video

Detects the subtitle language in the asset

To Add AI Checks in QA Checklist

  1. Open a new tab and navigate to Sprinklr Marketing > Advertising > Tools & Settings > Campaign QA Checklist.

  2. Open the checklist where you want to add the AI checks and click the + icon. To set up a new checklist, refer to How to Set Up Campaign QA Checklist.

  3. Hover over Ad Variant and select Define AI Checks from the dropdown menu.

  4. In the AI Checks pane, enter a relevant Check Name.

  5. Click Select Field to view and select the available AI check fields.

    Note: AI checks are separate from the usual QA checks due to the latency in the evaluation, while the usual QA checks run in real time. This latency occurs as our AI engine model evaluates the media asset frame by frame before scoring it.

  6. Click Select Operator to view and select the required operator for the check.

  7. Click Select Values to view and select the relevant value from the list.

  8. Click +Add Check to add additional AI Check fields.

View AI Checks Evaluation in Ads Composer

  1. Open a new tab and navigate to Sprinklr Marketing > Advertising > Ads Composer.

  2. Select the relevant Ad account and enter all campaign details to view the checklist evaluation on the right.

  3. Navigate to the Creatives tab and add a creative to view the AI check evaluations in the QA Checklist pane. All AI checks are grouped together and highlighted with the icon next to each check.

    Note: AI checks evaluation is not real-time and may take upto 1 minute.

View AI Checks Evaluation in Ads Manager

  1. Locate the required Paid Initiative in Ads Manager and click View & Edit.

  2. Click the checklist in the bottom right corner to view the AI checks in the checklist evaluation pane.

AI Check User Notifications

The platform will send you notifications whenever the AI Check evaluation is completed for your creative assets in the following scenarios:

  • Success in AI check evaluation: You will receive a notification when some or all AI checks for your media assets are successful.

  • Failure in AI check evaluation: You will be notified when all AI checks for your media assets fail.


FAQs

AI checks for your media assets may fail due to the following reasons:

  • File size: Images larger than 20MB and videos larger than 250MB are not supported.

  • ML Engine Timeout: If there are too many media assets in the creatives, the engine may struggle to evaluate them all. In the current evaluation architecture, all engines run in a queue and process data sequentially, without the ability to evaluate individual attributes in parallel.

  • Images: Less than 5 seconds for smaller images, and 15 seconds to 1 minute for larger images.

  • Videos: Up to 1 minute for videos shorter than 30 seconds, and up to 2 minutes for videos that are 1 minute long.