Demo Series Transcript: AI-Powered Recommendations for Hyper-Personalization
Speakers:
- Kristen Session, Product Marketing Manager, Blueshift
- Ajay Sundar, Senior Solutions Consultant, Blueshift
What Are Predictive Recommendations and Why Do They Matter?
Kristen Session: Welcome to today's demo on AI-powered recommendations for hyper-personalization. With the rise of AI tools like ChatGPT, marketers are looking for ways to use AI to deliver better customer experiences. Today, we’ll talk about how predictive recommendations help personalize what you show to customers.
In a world with endless options, personalization helps customers find what matters. Predictive recommendations use AI to match people with the products and content they are most likely to engage with. These recommendations adapt in real time based on behavior, improving conversions significantly.
For example, in our 2023 benchmark report, Blueshift customers saw a 166% increase in conversion rates with AI-powered recommendations.
Let me turn it over to AJ, who will walk you through the platform.
How to Build Smart Recommendations Using AI Recipes
Ajay Sundar: Thanks, Kristen. Let’s jump into Blueshift’s recommendations module. This is where you can quickly build personalized content blocks using over 100 pre-built AI recipes, tailored for industries like retail, media, finance, and more.
Let’s start by creating a block using one of the most common recipes: Abandoned Cart.
- Abandoned Cart Recipe: Shows items added to cart but not purchased.
- Requirements: Add-to-cart event and absence of purchase.
- Use case: Retarget customers to complete a purchase.
You can also:
- Add filters (e.g., only 4-star products)
- Exclude purchased items
- Control how many products show up
Now, let’s say you want to show similar items to what was abandoned. You can simply add a second block: Similar Items to Previous Block. This makes the email smarter by offering alternatives.
So just like that, you’ve built:
- One block for abandoned cart
- Another for similar items based on customer behavior
How Marketers Can Customize Recommendation Blocks
Each block gives marketers full flexibility:
- Inputs: Define events like add-to-cart, product views, or purchases.
- Filters: Target by brand, category, rating, etc.
- Outputs: Choose number of products shown or recommend at the category level.
It’s as easy as choosing a recipe and adding your custom logic—like cooking with a base recipe and personalizing with your own ingredients.
Using Recommendations Across Channels
Once built, your recommendation blocks can be used across:
- Email
- SMS
- Push notifications
- In-app messages
- Website
For example:
- In an email, you might show what Megan left in her cart.
- Below that, show similar items she might like.
- Or in push notifications, show abandoned items with just one click.
You build once and reuse across channels.
Testing and Proving Impact with A/B Campaigns
Once you’ve built your recommendations, you can test them. For example:
- A/B test emails with and without recommendations
- Measure engagement, clicks, and conversions
- Use performance data to optimize future campaigns
One customer saw a 166% lift in conversions from AI-powered product suggestions.
Real-World Customer Examples
Kristen Session: Let’s look at how real brands are using these recommendations:
Zumper:
- Uses real-time recommendations to match users to rental listings
- Scaled leads by 384% using automated, personalized campaigns
Skillshare:
- Uses recommendations to suggest personalized class content
- Their "personal picks" email saw a 71% higher CTR than standard newsletters
Final Thoughts: Smarter Marketing, Better Customer Experiences
With Blueshift’s AI-powered recommendations:
- Marketers get easy-to-use tools and 100+ recipes
- Customers get better experiences and faster paths to discovery
- Recommendations auto-update as customer behavior changes
From abandoned cart to cross-sell to price drops, the opportunities are endless.
To learn more, visit blueshift.com/contact-us. Thanks for joining us!
