The worst part of any new technology is the setup and implementation process. Since the early days of Hightouch, this is a problem we have been laser-focused on solving, and it’s the fundamental reason why we designed all of our products to be composable.
Our latest AI Decisioning platform is no different. Everyone wants to take advantage of agentic AI, but nobody wants to deal with the data infrastructure baggage that comes with it. We’ve intentionally designed AI Decisioning to work with any technology stack, regardless of whether you have a CDP, a Composable CDP, or no CDP.
Previously, the only way to use AI Decisioning was to integrate with your existing data platform. However, this new integration makes it possible for you to take advantage of AI Decisioning immediately, using Hightouch managed infrastructure. Companies can now choose between either bringing their own data platform or simply using Hightouch infrastructure to deliver richer 1:1 experiences to their customers.
Composability isn’t just part of our product, it’s in our DNA. In this blog post, we’ll show you how you can integrate your Segment data with AI Decisioning in 10 minutes.
Here’s a step-by-step guide for how this works:
Step 1: Connecting Segment to Hightouch
The first thing you have to do is configure your Segment instance to send data to Hightouch. To do this, you simply create a new webhook destination in the Segment UI and select the appropriate data sources (e.g., event streams) to send to Hightouch.
Note: this will handle all of your segment events, but we also have a range of SDKs and APIs we support that you can utilize if there are other data sources you want to leverage in AI Decisioning.
By default, Segment will automatically send Hightouch all the events, but you can easily constrain or limit which events you send using Segment’s mapping features. However, sending more data will almost always produce better results because it gives the system more context to train and fine-tune the agents toward your desired outcome. In fact, we actually recommend sending as many events as possible.
Step 2: Configuring the HTTP source in Hightouch
Next, switch over to the Hightouch UI. Navigate to Events > Event Sources > New HTTP Source. Hightouch will automatically generate a URL and a WriteKey.
Once you have this key, go back to Segment, and as part of the mapping process, input the URL you want to POST these events to and insert the WriteKey in the header.
Step 3: Testing and Configuration
Once you have taken care of all the preliminary steps, you’ll want to verify that your events are actually flowing from Segment to Hightouch. You can use Hightouch’s debugger to test these events. And after you’re satisfied that everything is properly configured, you can activate the integration. From that point on, your events will be fully connected to AI Decisioning.
Step 4: Building Marketing Flows
The final step is the fun part. Once your data has been forwarded to Hightouch, your marketers can actually start designing flows in AI Decisioning. Flows are individual use cases specific to your business and they’re unique to the outcome that you’re trying to drive. Your marketers maintain full control over the guardrails, creative, and channels. Instead of defining audiences and journeys, they just provide the inputs, and AI agents determine the optimal experience for every user.
The learning process can be really fast too. Within just six weeks of implementing the platform, WHOOP saw a 10% increase in cross-sell conversions, and PetSmart saw a similar result, driving a 22% increase in cross-sell campaigns. The system is also fully transparent, which means your marketing teams can view and understand every decision the platform is making for each individual customer.
Closing Thoughts
The beauty of this architecture is that it’s fully modular and composable. You can implement AI Decisioning today with or without a data platform or easily swap out the Hightouch managed infrastructure for your own data platform in the future. Best of all, swapping out the Hightouch infrastructure for your own data platform will not impact the algorithm or the learnings you previously accumulated.
Our goal with AI Decisioning is not to replace your existing MarTech stack or your marketing team. We want to supercharge both, so you can get more out of your investments and ultimately deliver better experiences to your customers—and this native integration with Segment makes this easier than ever.
