Accessible Intelligence – Marketing Streams

Learn more about Accessible Intelligence, and your potential use cases to launch.

Updating AI Marketing Streams

For each cluster, your AI marketing streams will specify, according to the algorithm:

  1. The number of emails within that stream
  2. The recommended values across key content categories for each individual email
  3. The wait time between delivery of each email

To view and update the content in your marketing streams, click on the Create Email Content button on the AI dashboard or Content tab on the left-hand navigation.

If more than one Use Case is available in your account, select the intended Use Case. Select the Edit Content option when hovering over the specific Cluster intended to work on. 

AI Edit Content Dashboard

Each email will specify if Awaiting Content, Needs More Work (with feedback from the classification team), or Approved.

Adding New Content

For an email Awaiting Content, select Add when hovering over the email placeholder. The content requirements will then be specified. You may choose to either:

  • Select an existing email, labeled by percentage match to the specified values. 

Note: The percentages are not weighted by importance of any individual category.

  • Start from scratch in building the email
  • Or, Filter based on the core criteria of the email, or the categories that would be most difficult to update from an existing email, such as the Taxonomy Object or Appeal Sub-Focus. When applying a filter, this will re-populate the email content available to select below, according to the filtered values.

AI Classification Filter

Selecting an individual email will allow you to edit the email content according to the specified values. 

Replacing or Editing Content

If selecting a specific email and hoping to further edit the content, change the base email, or update the email based on feedback from the content classification team, simply select and edit, or select to Update or Replace when hovering over the email thumbnail for the marketing stream.

Categories listed in the lefthand column in green are matched correctly. Categories listed in red require updating the email in order to match the recommended value. 

Hover over the ? icon next to any category to learn more about what it and the recommended value require.

Click and drag Text or Image blocks onto the email content to add, or edit existing content by hovering over the content block intended to edit, selecting the pencil icon, and using the WYSIWYG. Save any changes by selecting Insert.

Update the subject line and sender details via the Settings button at the top of the email. Click OK to save any changes. You may also choose to set an expiry date to remind team members to update the content after a certain period of time. This will NOT automatically pause delivery of the content and will continue to serve unless the marketing stream is ended or the content is updated and approved.

Email Settings

Submitting Content for Review

When ready to submit an email for review by the content classification team, click the Submit for Review button in the top right-hand corner of the email editor. Or, if not ready to submit, click Save to keep all updates made and revisit for later submission.

You will receive a notification based on the review of the email to alert if further updates are needed.

Marketing Stream FAQs

Q: How do the AI marketing streams impact how we should be communicating with supporters outside of clusters?

A: Marketing streams to clusters through AI can be considered as micro-marketing campaigns, in addition to the ongoing communication and standard campaigns that should continue for non-clustered supporters. 

A: No, testing is encouraged to continue in the ongoing marketing communications outside of AI to ensure a large enough audience to be statistically significant and a high enough volume of tests to ensure confidence and allow the AI algorithm to learn based on the results. As a general rule, the algorithm requires about 10 emails replicating the same test to be significant enough to factor into its recommendations.

Q: What happens if the algorithm recommends something we can’t update? What is the goal percentage to achieve as a match rate to the algorithm before submitting?

A: The recommendations are designed to achieve the maximum possible results, so the goal is always to achieve a 100% match rate. That said, circumstances change of certain email elements not able to be replicated based on campaign or organizational goals, etc. If you encounter a roadblock, please flag for the content classification team upon submitting. In some cases, they may share secondary values for next best recommendations for a particular category.

Try also using the content filters to identify the category values that may be more difficult to change and start with messages that match those values, such as filtering by taxonomy object, to update remaining values as needed, such as first emotion.

Updated on June 23, 2021

Was this article helpful?

Need More Help?
Can't find the answer you're looking for?
Contact Support