Introducing AI Decisioning

Since the early days of Hightouch, we’ve been laser-focused on empowering organizations to unlock the full potential of their data. Today, we’re announcing the next step in that journey: AI Decisioning.

The decisioning problem

Every business aspires to create more "super customers"—those who engage deeply with their products or services, contribute significant revenue, and remain loyal over time. But for marketing teams, making this vision a reality is easier said than done.

Each morning, marketers sit down to work and are confronted with a vast array of choices: Which customers to engage? What content will resonate? With which offer? In which channel? And at what moment? The vast number of options results in a constant struggle with difficult and often arbitrary decisions.

Tools like email, SMS, and push notifications can easily reach any customer, but how can a marketer know which combination of factors will actually influence customer behavior? With endless possibilities, no team, no matter how skilled or well-staffed, can manually test every option for every customer.

Methods like audience segmentation, journey building, and A/B testing attempt to bring some science to these questions, but they tend to fall short. In most companies, campaigns still feel generic, insights trickle in too slowly, and the guesswork never really goes away.

It’s a frustrating reality, but it sparked a thought— if humans alone can’t solve it, maybe humans and AI together can? That’s why we are thrilled to introduce our latest product: AI Decisioning.

AI Decisioning: The future of marketing

AI Decisioning optimizes your marketing programs to drive your desired customer behavior. For example,

• An airline may be highly motivated to drive high-margin upsells or cross-sells
• A retailer may want to drive customers to adopt in-store services that you’ve proven to increase repeat purchases and LTV (e.g., eye exams at your glasses store)
• A subscription-based app may want to drive engagement to prevent churn

Here’s how it works:

• Automated Experimentation: AI Decisioning constantly tests different combinations of messages, offers, and timing across your customer base to find the best possible way to interact with each of them.
• Individualized Decision-Making: Rather than manually defining rules around audiences, timing, or sequences, AI Decisioning analyzes all your data to determine the best treatment for each customer individually.
• Continuous Learning: With AI Decisioning, every decision is part of a continuous learning loop. The system doesn’t just make one decision and move on; it continuously makes predictions and experiments, learns from them, and adjusts its strategies to improve performance over time.

This powerful blend of automation, experimentation, and optimization allows your teams to focus on what they do best—strategizing and creating—while AI Decisioning handles the heavy lifting of experimentation, analysis, and decision-making for each customer.

Humans remain in control

AI Decisioning doesn't just magically run on top of all your marketing. Your teams are in full control to direct the AI - defining goals, crafting content, and establishing guardrails.

The first thing you need to do is create a "Flow" for any marketing initiative that you'd like to optimize. A flow has four main components:

1. Audience: The audience defines the span of customers you’re willing to target. For example, you may only want to drive cross-sells for healthy customers.
2. Content: Messages are the content you’re willing to send customers. The actual copy and creative live inside your engagement tool (e.g., Iterable, Salesforce Marketing Cloud, or Braze), but they’re referenced in Hightouch.
3. Outcomes: Outcomes define what the AI is optimizing for, such as purchases, email opens, clicks, etc. The framework around outcomes is sophisticated and supports multiple positive and negative indicators (e.g., unsubscribes) with custom or dynamic weights.
4. Guardrails: Guardrails provide constraints around who can receive what, frequency, timing, channels, and so forth.

A new product but the same Composable principles

We believe technology should be easy to adopt, quick to prove value, and transparent. Thats why we built AI Decisioning to follow the same core principles of the Composable CDP.

• AI Decisioning never stores or duplicates your data.
• AI Decisioning works with your existing data and schemas in whatever format they’re already in.
• AI Decisioning is a fully standalone product. It integrates with your existing martech, and you could even use it on top of a CDP other than Hightouch.
• AI Decisioning is not a “black box”. You can inspect exactly what is sent to each customer, as well as understand high-level trends and insights.
• You only pay for what you use. No shelfware.

What makes AI Decisioning different?

People have been talking about sending the ‘right message, to the right person, at the right time’ forever. And more recently, there’s been an influx of shiny new AI products.

Today, most AI marketing solutions are focused on using machine learning to predict the likelihood of customers taking certain actions. This is often referred to as “next best action” or “predictive modeling.” Having a predictive score or “next best action” still leaves a majority of the hardest decisions completely unsolved: What’s the best message to actually drive that action? Which variation will resonate with this customer? When and how often should it be sent? On what channels?

AI Decisioning uses continuous experimentation to test all of these decisions in conjunction, rather than producing a static prediction and leaving it to the marketer to figure out every other decision through manual rules and testing.

Ready to learn more? Book a demo to see AI Decisioning in action and discuss real use cases.

We hope you’ll join us as a customer, partner, friend, or just supporter in building a better future for marketing together.