Welcome to part three of our four-part blog series on our AI study with Wakefield Research and Ad Age. In previous installments, we’ve covered what marketers really think about generative AI and how AI automation and optimization impacts today’s marketers.
Today, we’re talking about what Ad Age refers to as “the foundation and future of all AI marketing:” AI explainability and optionality and Predictive Goals. These aspects of AI marketing are why Iterable specifically is at the forefront of AI innovation.
So let’s jump in.
Want to read through all our AI research? Head to Ad Age now to download the full study.
What are AI Explainability and Optionality?
As Ad Age states, AI optionality and explainability are “the next stage of evolution for automation […] in which marketers can take even more cues from AI.” With this two-pronged approach, “the AI not only automates the deployment of a campaign, but it also automatically generates the campaign itself.”
If the idea of an AI deciding which of your audiences to target—and how—for example, makes you apprehensive, that concern is valid. That’s why explainability and optionality are so crucial to AI marketing:
- Explainability establishes trust by infusing transparency into the AI model’s decision-making process
- Optionality provides marketers the option to override the system’s recommendations
In a previous blog post about key AI terms to familiarize yourself with, we further defined Explainable AI:
“Explainable AI involves transparent systems with clear, understandable processes. Unlike the opaque, black box of certain AI solutions, Explainable AI provides a more ‘glass box’ experience that shares deeper insights into the data that powers predictions.”
Let’s now walk through an example to demonstrate explainable AI in action.
Explainable AI in Action
Ad Age illustrates the concept of explainable AI with an example—a fitness chain. To achieve its goal of converting free mobile app users to paid subscribers (a.k.a. “freemium” to “premium”), the business could use an AI model to target people who have signed up for three fitness classes. AI has identified that this particular threshold has a high statistical correlation to the campaign’s desired outcome, and explainability is how the algorithm shows its work, so to speak.
In this example, with the right marketing platform, marketers can pull up a dashboard with details that reveal how the model arrived at its conclusions. That way, the team can be reassured that the AI is working effectively—before they hit the send button.
Adriana Gil Miner, CMO of Iterable, discusses the reasoning behind this explainable AI:
“For instance, you will be able to see that the people most likely to move to the paid app watched three videos and logged in five times in 30 days. This builds transparency and trust. It validates the audience and the quality of the prediction, and further, it gives marketers insights on how to drive their business with ideas for other strategies.”
Explainability and optionality were built into Iterable’s AI from the outset, so let’s dive into how marketers can take advantage of explainable AI with our suite of tools.
AI Explainability With Iterable
Iterable’s AI Suite was designed to empower your marketing and forge deeper customer connections, and at its core is putting people in the driver’s seat of AI-assisted decision-making.
This built-in intelligence uses explainable AI at the heart of several features:
You can click through to get a comprehensive explainer of each feature, but here’s the TL;DR.
Brand Affinity
Brand Affinity harnesses Iterable AI to label customers based on their level of engagement with your brand. You can use these labels to provide rewards to loyal users, improve open rates by suppressing negative users, test offers by sentiment, and so much more.
If you’re wondering how these labels are assigned, that’s where Iterable’s Explainable AI comes in. Simply click into a specific user’s Brand Affinity score to view additional insights, such as:
- A user’s affinity labels for the past 30 days to see what changes may have occurred.
- A summary from the past 90 days of the campaigns that contributed most to a user’s past and current affinity scores.
- A summary from the past 90 days of the metrics that contributed most to a user’s affinity score with an explanation of how those actions compare to the behavior of users.
Iterable’s Explainable AI breaks down the contributing factors to each user’s Brand Affinity score.
It’s not enough for an AI model to score subscribers by sentiment. It must be able to explain why it’s given those scores and how these factors impact overall engagement.
Next Best Action
As Ad Age stated, we’re now entering a technologically advanced era in which an AI model can not only deploy a campaign, but also generate it. Iterable’s Next Best Action takes this to the next level, analyzing the end-to-end logistics of your campaigns and generating audience and copy recommendations to improve areas of performance.
With just a few clicks, you can automatically create a campaign, select the right segment and come up with the perfect subject line. You’re still in full control, but now you can access an AI-assisted push in the best direction.
Automatically send campaigns to the right customer segments using AI-optimized subject lines with Iterable’s Next Best Action.
Predictive Goals
In the last installment of our AI blog series, we’ll be covering Predictive Goals in more depth, but for the sake of this post, Iterable uses Explainable AI within its Predictive Goals feature to analyze your historical data and predict which users are most likely to convert on your business goals in the future.
Like Brand Affinity, Predictive Goals includes modules that explain how to understand the predictions the AI model is making. The platform lists the custom events and user fields that are contributing to the prediction, and surfacing which data points are making the most meaningful impact.
As seen in the image below, you’ll see a predictive strength score and an interactive view of your prediction results, including the events and properties that Explainable AI indicates have the greatest potential to influence your outcomes.
The module on the left-hand side in Iterable’s Predictive Goals displays how Explainable AI is reaching its prediction conclusions.
And when you click the “Explore” button to learn more, you’ll see a panel like the one below that shows the number of events and properties that Predictive Goals evaluated when the prediction was generated. It also displays a breakdown of the statistically significant contributors.
Clearly see exactly what factors are shaping your Predictive Goals with Iterable’s Explainable AI.
With Iterable, nothing is a mystery. While its AI-driven insights are sure to be illuminating, you’ll never be stuck wondering how the model is making its decisions. That’s our “glass box” promise.
Let Us Explain Iterable’s Explainable AI
We have so much more to cover on Predictive AI, so stay tuned for the final blog post in this AI series.
You can learn about all of the features we discussed and more in Iterable’s AI Suite. And don’t forget to download the full Ad Age report for even more insights about the next evolution of AI technology.
Ready to get even more explanation about how to advance your AI marketing? Reach out and schedule a custom demo of Iterable AI today.