Act-On’s AI Audience Insights feature uses a machine learning model to analyze contact behaviors and profile data and combine “clusters” of contacts that are similar in some ways.
How Does It Work?
The model analyzes standard profile features such as job title, industry, state, and country. It also analyzes over 40 behavioral data points, such as the # of emails opened, # of forms submitted, # of webinars registered, etc. It looks at these behaviors in the last 30, 90, and 180 days and can also analyze how these behaviors may change over time.
With all this data, the model is able to group contacts together based on similarities that may not be immediately obvious to us humans. Once it creates these groups, we sort through the recommendations and prioritize them. The most potentially relevant recommendations are presented as suggestions on the Act-On Insights page.
Access these suggestions from the 3rd data card at the top of the All Contacts page, or from the dropdown menu in the upper right corner.
What Can I Do With These Suggestions?
If you see a suggestion that seems interesting, you can create a segment from it. This will translate the recommendation into a query-based segment of your All Contacts list. From here, you can use it like any other segment. Send an email, use it in an automated program, or use it as a suppression list.
If you don’t think a suggestion is relevant to you, discard it. This will remove it from the recommendations list and let us know that it wasn’t useful. You may be asked for further feedback as to why you chose to discard it. Accepting and discarding segments will provide further data to refine the model and make it more relevant to you in the future.
How Can I Customize These Recommendations?
The model runs on a large number of behaviors and profile fields. You may find that some of them are not relevant to your business or your contacts. Click the Customize button to exclude these from the model’s calculation. You can also choose to limit the time frame for behaviors to be included.
If you’d like to limit the contacts included in the recommendations, you can also do this via the Customize button. For example, you could choose to focus only on leads instead of all contacts. This may give you more significant or relevant results.
When you save your customizations, the model will apply these settings the next time it runs (about once per 24 hours).
How is My Data Used Or Shared?
Your data is not shared outside your account and is not used for training purposes. The model is running statistical analysis on your account’s CRM contact data and their associated behaviors to extract correlations. These are displayed only within your account. Act-On uses Chat GPT to review and summarize the recommended segments, but it does not have access to the raw data.
What is Required to Use This Feature?
This feature is only available for CRM-synced contact data. Also, your account must be on Act-On Contacts to see the recommendations and save the recommended segments.
If your account qualifies, you can activate this feature in Labs. Go to Settings > Other Settings > Custom Account Settings > Labs. You must be an admin user to activate this feature.