Hightouch launches AI Decisioning

Composable CDP Hightouch announced a new enterprise AI product, AI Decisioning, as a native Snowflake app. AI Decisioning leverages Snowflake Cortex AI, a suite of tools that runs large language models against unstructured data.

AI Decisioning continuously experiments with customer data available in Snowflake to identify the best ways to engage with individual customers based on any attribute or behavior. It integrates with platforms like Salesforce Marketing Cloud, Braze and Iterable for activation. It is intended to move beyond targeted segments to suggest engagement events for individual customers.

“When marketers build audiences for campaigns, they inherently lump together many individuals based on overly broad generalizations and their best guess on customer behavior. AI Decisioning removes the guesswork.”

Brian Kotlyar, VP of marketing and growth at Hightouch (in a release)

Why we care. On the one hand, this presents as an ambitious attempt at automating analysis of data throughout the warehouse/lake at the level of the individual customer. What is less clear is how it is to be differentiated from AI-powered next-best-action recommendations for individual customers which has been available for some time, for example from Pega, unless it’s a broader scope and larger scale of data being analysed.

A composable approach. We asked Brian Kotlyar, VP of marketing and growth at Hightouch, about what differentiates this product. “As with all of our products, Hightouch comes at this problem with a composable approach,” he said. “That means things like never storing or duplicating data, working with existing data and schemas, integrating with existing martech, and only asking customers to pay for what they need and use.”

In essence, instead of turning a data lake into a composable CDP, as Hightouch’s existing offering does, the new product turns a data warehouse or data lake into an AI Decisioning hub. “That’s really attractive to most enterprises,” said Kotlyar.

MarTech contributor Craig Howard, chief solutions officer at customer data accelerator Actable, had a positive response to the announcement. “I think what Hightouch is doing is an excellent addition to their functionality. They have a been a fast innovator [in defining] what a ‘composable CDP’ can be. This is another step towards expanding their orchestration capabilities, whereas other providers have focused more towards data management and governance.”

On the comparison with Pega, he said: “It will take time for them to reach the maturity and scale of a purpose-built decision management platform like Pega (which can decide on thousands of pricing and content options across channels), or even something like Optimove, where decision management is at the core of their orchestration approach, however this is great entry-level decision management capability for Hightouch.”

The success of the solution will depend, he said, on how well you organize the data and the quality of that data. This is a hurdle for any solution relying on a data warehouse: “The devil is in the details. Consistent and meaningful metadata is key to this particular functionality.”

MarTech contributor Greg Krehbiel wondered whether there might be a disconnect between the segments discovered and the ability to act on them. “On the other hand,” he said, “AI could discover groups that the marketing department is unaware of and that could be used to decide where to focus efforts.” Echoing Howard, he also said: “The devil will be in the details.”

Human in the loop. If the idea of a powerful AI solution experimenting continuously behind the scenes and popping out engagement suggestions for your customers is daunting, be assured that there remains a human in the loop. Human marketers set the desired goals and outcomes; human data teams manage the data that is accessed.

Dig deeper: What the composability revolution means for CDPs