Segment's Q2 product updates focus on enhancing AI-driven personalization with property-level predictions, product-based audience recommendations, and generative text-prompt audience creation.
The Twilio Segment products introduced features around three key themes in Q2 of this year - warehouse interoperability, AI and personalization, and Platform Extensibility.
We’re recapping the latest updates in this blog series, where we started with a blog covering how Segment helps customers unlock new personalization use cases with data flowing between the data warehouse and CDP.
Keeping up with personalized content, this post will serve the feature themes of Q2- AI and Personalization and Platform Extensibility!
Last year, we launched CustomerAI Predictions, which lets marketers use the power of AI to predict the likelihood of their customers to perform any tracked event in Segment. Once those Predictive Traits are created, we offer up pre-built Predictive Audience segments to help inspire new campaign ideas and activate them quickly to drive desired business results without needing to rely on data science resources.
This quarter, we’ve launched some new functionality within Predictions with the ability to predict on the property-level. This has been a highly requested functionality that is now generally available to all Unify Plus and Twilio Engage customers.
For example, you can now predict purchases where the cart size is more than $50, instead of just predicting the likelihood of someone to purchase or add to cart.
So you make much more detailed and targeted Predictions which will be exciting for lots of customers who have been looking for this level of flexibility.
Where CustomerAI Predictions solves “who” to target, Recommendations solves “what” to target them with.
In our current private beta, we’re enabling customers to build product-based audiences. We're making it simple to automatically determine which specific product, brand, or product category your customers are most likely to purchase so you can then send those targeted audiences to downstream channels like ad platforms, or launch multichannel customer journeys.
Consider a use case where you’re a retailer and you need to offload some excess inventory after a big seasonal event. If, for example, you have too many Adidas sneakers in stock, you could use Recommendations to create an audience of customers who are most likely to buy Adidas sneakers and send them targeted promotions instead of slashing prices across the board for those items.
This new CustomerAI feature can help you quickly provide customers with more relevant and personalized experiences that boost conversions and grow average order value.
This feature is currently in private beta and will go into general availability in Q3.
This is the first generative feature we’ll be launching, empowering marketers to create target audiences in minutes with a simple text prompt.
Instead of going through every stage of our audience builder in Twilio Engage, you can type out the specific audience you want to build into a text box, and Engage will automatically generate that audience for you.
For example, you can write: “Create an audience of customers who have spent at least $50 in the last year and have visited our website but not completed a purchase in the last 30 days.”
This will save marketers time, boost productivity, and reduce time to launch for campaigns.
This feature is currently in private beta.
This new feature helps support more streamlined and enriched transactional-style journeys in Engage that consist of sending users communications after they take an action. So you can think of use cases like:
Order confirmations
Welcome messages
Event registration confirmations
Historically, we’ve sent events from Source to Destinations via Connections and that works well, but the events sent via Connections are not associated with the profile and they don’t go through identity resolution. And this is important, contextual information marketers need to better understand and segment their audiences in their downstream third-party tools like Braze, Iterable, or Customer.io.
Now, with Event Triggered Payloads, you can create more simplified transactional journeys and the payload that is sent to your third-party tools will include all the context and trait properties from the event associated with the identity-resolved profile.
For example, if a customer abandoned a journey, why did they abandon it? Maybe their size or color preference wasn’t available. With Event Triggered Payloads, you’d have that information in Braze or Iterable to help you further segment and personalize your communications to that person.
This is in Private Beta this quarter.
We’re thrilled to announce that we are extending the power of CustomerAI to developers with Functions Co-Pilot, powered by OpenAI.
Functions Co-Pilot consists of two features that let developers make use of generative AI to extend what is possible with Segment.
The first, Functions Writer, generates code from natural language prompts. In other words, it turns what would previously have been weeks of integration coding into a simple prompt that’s about the same length as a text message or a Tweet.
Last quarter, we added rETL support for Functions. This means you will even be able to use Functions Writer to generate rETL flows to send data from the data warehouse to custom destinations.
For existing code, Functions Analyzer has you covered. Functions Analyzer scans your code, summarizes what it does back to you, and then offers recommendations for how you can optimize the Function.
The recommendations you receive from Functions Writer and Analyzer are based on proven methodologies from experienced subject matter experts.The underlying model combines context from your current workspace configuration, Segment’s vast library of example Functions and code templates, and general best practices for Javascript to arrive at the optimal solution for your use-case.
Think about all of the powerful tasks a Function can do - from integrating custom sources and destinations (including the data warehouse), enriching data with third party sources, transformations, and even tokenizing/encrypting/decrypting data before it reaches downstream destinations. Now all you need is a couple of sentences and your trusty AI guide to unlock endless innovative, cutting edge use-cases that push the boundaries of what is possible with a CDP.
Functions Co-Pilot launches to public beta in Q2.
Helping our customers maintain high quality data is Segment’s raison d’entre, so it comes as no surprise that we’re launcing a new feature designed to help you enhance the overall quality of your profiles.
Unify Event Filtering extends the controls and customization features that Segment provides to users by allowing you to block events from entering Unify spaces. The events are blocked prior to identity resolution, so they will not be sent to profiles. This feature supports filtering for all event types and by event name, properties, and traits. If you have separate Unify spaces that cover different regions and different product lines, it can be helpful to isolate these spaces by configuring different controls for what events go into them.
Unify Event Filtering enables you to curate profiles so that your marketing and product teams are only activating relevant profiles with high quality events when delivering personalized customer engagement. This also streamlines workspace configuration as you don’t have to configure your Connections workspace with different Sources to reroute events, accelerating time to value and reducing MTU consumption from the previous workarounds.
Unify Event Filtering will be launching to Private Beta in Q2.
In Q2, our focus was on enhancing AI-driven personalization and platform extensibility. We introduced features like property-level predictions, product-based audience recommendations, and text-prompt audience creation. Additionally, the new Event Triggered Payloads feature supports enriched transactional journeys by sending context and trait properties from identity-resolved profiles to third-party tools.
Our annual look at how attitudes, preferences, and experiences with personalization have evolved over the past year.