Introducing Insights for AI Decisioning: Uncover hidden marketing opportunities

Your cross-sell email campaign has been running for six weeks. What have you actually learned?

Maybe conversions are up. Your new subject line is performing better. But why? Who’s responding? What’s really driving the lift?

Finding these answers today is a guessing game for marketers. They dig through dashboards, slice data by demographics, test different segments, and try to connect the dots—only to realize they’re working with limited variables and superficial learnings. The most interesting questions go unanswered.

Today, we’re launching a new part of our AI Decisioning platform called Insights. Insights uncovers unexpected correlations and hidden knowledge you never could’ve found before by hunting across vast pools of data created by your AI agents as they experiment.

It automatically surfaces significant patterns and opportunities that you would struggle to find manually, so you can better understand which creative, offers, messages, and schedules resonate with which parts of your customer base.

With Insights, marketers can learn alongside AI, finding and capturing new ideas that marketers can use across marketing channels. Let’s explore how it works.

You’re hunting for learnings, but what if the insights found you instead?

In most marketing, how much you learn is simply a function of how many iterations of a campaign you can run and analyze. Today, lifecycle marketing is mostly based on running a limited number of broad campaigns across a limited number of rigid pre-defined segments. That means that the opportunities to learn and the conclusions you can draw are … limited.

To make this real, let’s imagine a hypothetical jewelry retailer. They might send a campaign to women over 40 who have previously purchased at least one item of jewelry and they might do it once per week. That gives our jewelry retailer one chance to learn about one audience every week—if they are diligent at A/B testing, perhaps a little bit more. With this one chance, the marketers will certainly learn something, but it’s going to be slow, and it’s very rare to uncover anything truly new or insightful.

AI Decisioning works very differently. With AI Decisioning, you simply set a goal and a set of actions that your AI agents are permitted to perform and then step back as your agents test everything.

AI Decisioning creates learning opportunities beyond traditional marketing.

This “test everything” approach creates a massive new surface area for unexpected correlations and interactions to emerge. The problem of learning stops being about planning and measuring a specific test, and much more about finding valuable insights in the massive pool of data created by your AI agents. We built our latest Insights capabilities to find and surface those novel correlations and interactions for marketers clearly and simply.

For contrast, let’s compare our theoretical jewelry retailer with what really happened at WHOOP in a short time with AI Decisioning.

Answering questions they didn’t know to ask: WHOOP’s journey with Insights

WHOOP, a wearable health and fitness company, used AI Decisioning to drive 10% lift in cross-sell conversions through email campaigns. What they were most surprised by, though, was the value of insights that emerged incredibly quickly as AI agents optimized their emails.

As AI Decisioning tested a range of activity-oriented messages, it uncovered unexpected results. For example, users whose top activity was swimming clicked through on martial arts content at a higher rate than even users who listed martial arts as their top activity.

AI Decisioning led WHOOP to new insights about swimmers and martial arts.

This is just one example of WHOOP’s learnings. Among other highlights, they also learned how different markets prefer different days of the week for messages, and how different propensity groups respond to significantly different messages.

These weren’t discoveries WHOOP was actively looking for, but once uncovered, they opened new doors. The company is able to use these learnings not just in AI Decisioning campaigns but across all marketing channels, from email to ads to in-app engagement.

Feature deep dive: Insights uncovers marketing opportunities

The power of AI Decisioning is as much in the findings it generates as its direct campaign optimization. We generate two types of insights automatically:

• Top performers: Your best-performing user segments and creative variants
• Personalization insights: Your best-performing actions (such as messages and timing decisions) for specific user segments

For both types, Insights algorithmically combs through thousands of charts and comparisons and then surfaces only the most statistically significant or anomalous results. Let’s look at what you can do with each type of insight.

Top performers: What’s working at a global level

Top performers shows your best-performing user segments and creative variants.

Top performers breaks down your campaign performance across your entire AI Decisioning audience by a single user or action dimension, like age, propensity scores, or message themes.

These insights answer global questions, such as:

• How does the conversion rate differ by state?
• Does a customer’s predicted propensity to buy actually impact campaign performance?
• What are our best-performing subject lines?

Insights highlights the particularly significant or anomalous results because these discoveries are often the most actionable. When one subject line clearly outperforms others, you can dig deeper to understand why, such as through customer interviews. You can then decide if you should focus more on that message. Similarly, if a single day of the week is best for emails, you can test whether that’s true for other channels.

Personalization insights: What’s working in micro-context

Personalization insights shows your best-performing actions for specific user segments.

Personalization insights add an additional dimension—it surfaces patterns and correlations across your message and users. With personalization insights, you can answer these types of questions:

• How do customers in Illinois respond to my subject line variations compared to all users?
• What time of day do customers in the 20-29 age range engage with email?
• Which channels have the highest conversion rate for customers with a high buying propensity?

Each question hits on a particular action and a segment of the AI Decisioning audience. What makes this type of insight so valuable is the comparison against the entire audience. That Californians respond to text messages more regularly than emails or push messages is interesting only if that significantly differs from the entire audience of the campaign. Once that’s determined, your marketing team can create more geographically distinct messaging.

On-demand performance analytics

Charts show your campaign performance against holdout and control groups.

Insights builds on our existing campaign performance charts, which enable marketers to self-serve charts to understand AI Decisioning campaign performance in relation to holdout groups. Marketers can segment and group performance data by any available customer data properties. When you’ve built the chart you need—or that helps you understand overall performance for a segment identified by Insights—save it for regular use.

Powerful learning for marketer success

Insights for AI Decisioning creates something remarkable: automated discovery that surfaces patterns and opportunities you would struggle to find manually. As AI Decisioning runs thousands of experiments while optimizing your campaigns, Insights captures and translates learnings into intelligence that you can actually use.

This creates a relationship where AI and marketers grow smarter together:

AI learns to optimize your campaigns: Each decision makes our AI agents more sophisticated in understanding customers and refines their approach based on real data.

Marketers gain unprecedented visibility: You see what’s working, with whom, and why. These findings uncover customer patterns and preferences that would remain hidden without the experimentation scale of AI Decisioning, building a foundation for more informed strategic decisions.

The impact compounds across your business as the patterns you discover reveal new market opportunities and highlight untapped segments worth pursuing in additional channels.

Insights supercharges your marketing intelligence. Learning alongside AI Decisioning turns data into understanding, understanding into action, and action into incredible results.

We hope you’ll give it a try. Contact our solutions team to learn more.