How E-Commerce Marketers Use Klaviyo's Segments AI to Build Complex Customer Audiences

Customer segmentation is the engine behind high-converting e-commerce lifecycle marketing. Moving beyond generic email blasts to hyper-targeted messaging consistently drives higher open rates, lower unsubscribe counts, and measurable lifts in revenue per recipient. However, manually assembling multi-condition audiences in traditional segmentation builders has historically been tedious and prone to human error. Marketers had to click through nested dropdowns, construct elaborate Boolean statements, link disparate behavioral events with "AND" versus "OR" connectors, and verify metric lookback windows.

Klaviyo’s Segments AI—anchored by the "Define with AI" workflow—streamlines this entire process. Powered by natural language processing and generative AI, it allows e-commerce brands to describe their ideal audience in everyday language and automatically translates those descriptions into production-ready segment criteria. Here is a comprehensive guide to understanding how Segments AI works, how to structure high-performing prompts, and how to deploy complex audiences across email and SMS workflows.

What Is Klaviyo Segments AI?

Segments AI is an assistive intelligence capability embedded directly within Klaviyo’s segment builder and campaign creation flows. Instead of requiring marketers to configure individual conditions one by one, the tool parses plain-text prompts—sentences or descriptive phrases—and converts them into structured Klaviyo segmentation logic.

Under the hood, Klaviyo maps conversational intent directly against:

• Behavioral metrics: Actions tracked via native e-commerce integrations like Shopify, WooCommerce, and BigCommerce (e.g., Placed Order, Viewed Product, Started Checkout).
• Profile properties and custom attributes: Stored contact details such as geographic location, customer tags, loyalty tiers, or custom form submissions.
• Predictive analytics: Machine-learning forecasts, including customer lifetime value (CLV), churn risk, and expected date of next order.
• Engagement history: Email opens, SMS clicks, and subscription timelines.

Audiences created with Segments AI retain Klaviyo's signature dynamic behavior. They are not static email lists. Once created, profiles automatically enter or exit the segment in real time as their customer attributes and event streams change. Additionally, Klaviyo tags every audience created via this tool with a dedicated klaviyo-ai tag, making it easy to audit and filter AI-generated segments across your account.

Where to Access "Define with AI" in Klaviyo

Marketers can launch Segments AI through two distinct entry points in their Klaviyo account (available on paid plans):

1. The Lists & Segments Menu

1. In your Klaviyo dashboard, navigate to Audience > Lists & segments.
2. Click Create New in the top right corner and select Create segment.
3. Name your segment.
4. Click the Define with AI button situated near the condition selector (identifiable by the AI sparkle symbol).
5. Type your prompt into the text field and press Enter or click the suggestion prompt.
6. Review the generated conditions, click Open in builder to make manual refinements if needed, and click Create segment.

Important Workflow Rule: Segments AI is designed exclusively for building new segments from a clean slate. It is not available when editing existing segments, and triggering it will overwrite any conditions you have already entered manually into the builder.

2. The Campaign Wizard Flow

Marketers do not have to leave their campaign creation process to assemble new targeted groups:

1. In the campaign builder, proceed to the Recipients step.
2. Click into the Send to or Don't send to selection box.
3. Type a natural language description of your target group instead of picking an existing list.
4. Select Generate new segment at the bottom of the dropdown list.
5. Preview the resulting logic, refine it if necessary, and attach the newly created segment directly to the campaign send.

The Anatomy of an Effective Segment Prompt

While Segments AI can interpret basic descriptions like "engaged subscribers in the last 30 days," building nuanced, revenue-generating e-commerce audiences requires intentional prompt structure. A complete, production-grade prompt consists of four essential elements:

The Critical Rule: Always Specify Marketing Consent

By default, generative AI does not automatically add channel consent criteria. If you type "people who bought running shoes in the last 30 days," the AI will construct conditions based purely on purchase history. This segment could inadvertently pull in customers who purchased as guests or previously unsubscribed from promotional marketing. To safeguard message deliverability and adhere to privacy regulations, always include explicit statements like "profiles consented to receive email marketing" or "consented to SMS marketing" directly within your prompt.

4 Advanced E-Commerce Audiences Built with Natural Language

Here is how growth marketers translate complex e-commerce revenue strategies into prompt-driven segments:

1. High-Intent Window Shoppers (Browse Abandonment)

• Marketing Objective: Retarget store visitors who showed clear purchasing intent across recent sessions but stalled before buying, offering category-specific recommendations or low-friction incentives.
• Natural Language Prompt: Profiles consented to receive email marketing who have viewed a product at least 2 times in the last 14 days, have not placed an order in the last 30 days, and have not started checkout in the last 7 days.
• Generated Logic:
  • Can receive email marketing = True
  • AND Metric: Viewed Product at least 2 times in the last 14 days
  • AND Metric: Placed Order zero times in the last 30 days
  • AND Metric: Started Checkout zero times in the last 7 days

2. High-LTV VIPs (Retention & Loyalty Upsell)

• Marketing Objective: Identify your top brand advocates for early access to product drops, exclusive VIP sales, or loyalty club perks.
• Natural Language Prompt: Profiles consented to email marketing who have placed an order 3 or more times over all time, with total order value greater than $300, and have opened an email in the last 45 days.
• Generated Logic:
  • Can receive email marketing = True
  • AND Metric: Placed Order >= 3 over all time
  • AND Metric: Placed Order Total Value > 300 over all time
  • AND Metric: Opened Email at least once in the last 45 days

3. At-Risk Churn Prevention (Win-Back Cohort)

• Marketing Objective: Intercept previous buyers who have exceeded their typical reorder timeline before they lapse permanently.
• Natural Language Prompt: Customers who placed an order between 60 and 180 days ago, have not placed an order in the last 60 days, and are consented to receive email marketing.
• Generated Logic:
  • Can receive email marketing = True
  • AND Metric: Placed Order at least once between 60 and 180 days ago
  • AND Metric: Placed Order zero times in the last 60 days

4. Cross-Channel Engaged VIPs (SMS + Email Synced)

• Marketing Objective: Identify multi-channel subscribers who actively interact with both messaging channels to drive flash sales with maximum engagement.
• Natural Language Prompt: Profiles consented to email and consented to SMS who have opened an email or clicked an SMS in the last 30 days and have purchased at least once in the last 90 days.
• Generated Logic:
  • Can receive email marketing = True
  • AND Can receive SMS marketing = True
  • AND (Opened Email >= 1 in last 30 days OR Clicked SMS >= 1 in last 30 days)
  • AND Metric: Placed Order >= 1 in last 90 days

Best Practices for Review and Optimization

Klaviyo’s AI generates an accurate, logical baseline, but e-commerce marketers retain full responsibility for campaign compliance and deliverability. Follow these best practices to ensure seamless execution:

1. Conduct a Logic and Operator Audit: Always inspect the generated segment rules in the visual builder before hitting publish. Verify whether the AI used "AND" (every condition required) versus "OR" (any condition valid) appropriately, especially when bridging email activity and SMS behavior.
2. Confirm Store-Specific Event Names: If your store utilizes custom integration events or headless checkout webhooks rather than standard platform metrics, verify that the AI selected the correct event definition (for instance, custom checkout events versus standard Shopify "Placed Order").
3. Layer Predictive Analytics Manually When Needed: While natural language prompts can isolate behavioral thresholds, you can click Open in builder to layer advanced Klaviyo predictive fields—such as filtering for Predicted Churn Risk is High or Expected Date of Next Order is in the next 14 days—to build predictive replenishment flows.
4. Use AI Segments to Trigger Automated Flows: Segments AI is not limited to one-off broadcast campaigns. Because Klaviyo segments calculate dynamically, you can use these audience groups as entry triggers for customized automated flows or as segment filters to bifurcate existing customer journeys.

By combining plain-text prompting with structured human oversight, e-commerce marketing teams can eliminate manual configuration friction, launch sophisticated micro-segments in minutes, and maintain high personalization standards across every customer touchpoint.

References

• Customer Segmentation Tools for Smarter Audience Building - Klaviyo
• AI Workflow Automation Tools for Better Marketing & Customer Service - Klaviyo
• How to define segments with AI | Klaviyo Help Center
• Advanced segmentation reference | Klaviyo Help Center
• Define segments with AI | Klaviyo Academy
• Segments AI - Klaviyo