Ecommerce

AI for abandoned cart recovery: real numbers from stores

20 April 2026

What is AI for abandoned cart recovery?

AI for abandoned cart recovery refers to the use of artificial intelligence systems to identify customers who have added products to a cart without completing checkout, and to reach them with personalised follow-up messages designed to bring them back to complete the purchase. This goes beyond standard abandoned cart email sequences by using conversational AI (chatbots via WhatsApp or SMS) to personalise the follow-up to the specific products the customer left, handle objections in real time, and apply timing and channel logic based on the individual session rather than a fixed schedule.

What do the recovery numbers actually look like?

Standard abandoned cart email sequences recover between 5 and 8% of abandoned sessions. This figure varies by product category, price point, and the quality of the email sequence, but 5 to 8% is the realistic range for a store running a standard three-email sequence. AI-driven conversational recovery via WhatsApp, where the store has customer contact details and opt-in permission, has recovered 11 to 14% of abandoned sessions in stores where it has been properly configured and run for 30 or more days.

The gap between 6% and 12% sounds modest in percentage terms. In revenue terms it is not. A store doing 200,000 per month with a 70% cart abandonment rate and a 2% baseline conversion rate has approximately 14,000 carts abandoned per month. Recovering 6% of those returns 840 carts. Recovering 12% returns 1,680 carts. At a 60 average order value, the difference is 50,400 per month in recovered revenue, not recovered from new traffic but from traffic the store already paid for. That number is why AI cart recovery commands the attention it does among ecommerce operators.

What conditions produce these recovery rates?

The 11 to 14% recovery rates come from stores that meet specific conditions, and the numbers drop significantly when these conditions are not present. The first condition is having the customer contact details for the abandoning session. Most cart abandonment happens from users who are not logged in and have not provided an email or phone number. The follow-up message has to go somewhere. Stores that have implemented email capture in the add-to-cart flow, or that use returning customer recognition to connect anonymous sessions to known customers, have a larger addressable pool than stores that only recover from users who previously created an account.

The second condition is permission to message via WhatsApp or SMS. WhatsApp open rates run 70 to 90% for business messages. Email open rates for abandoned cart sequences run 20 to 25% on average. The channel multiplier is real, but WhatsApp Business messaging requires explicit opt-in. Stores that have built WhatsApp opt-in into their checkout flow, loyalty program, or post-purchase sequence are positioned to use the channel. Stores that have not cannot. The investment in building WhatsApp opt-in before running cart recovery is the prerequisite investment that determines the ceiling on recovery rates.

The third condition is chatbot configuration quality. A chatbot that sends a generic message about your cart is waiting recover 5 to 6% at best. A chatbot configured to reference the specific products left, to lead with the most common objection for that product category (size uncertainty for fashion, ingredient questions for supplements, compatibility questions for electronics), and to offer a relevant pathway to resolution recovers 3 to 4 percentage points more than the generic version. That configuration work is 4 to 8 hours per product category. It is the setup time most stores skip, and skipping it is why the tool never performs to the case study benchmarks.

What channels does AI use for cart recovery?

The channels in use for AI cart recovery in 2026 are, in order of effectiveness: WhatsApp (70 to 90% open rate), SMS (60 to 70% open rate), email (20 to 25% open rate), and browser push notifications (3 to 5% open rate). WhatsApp delivers the highest recovery rates because the open rate is higher, the response rate is higher, and the conversational format allows real-time objection handling. SMS delivers good open rates but lacks the conversational interface. Email is the lowest-barrier channel to implement but the lowest-performing for recovery in 2026.

The optimal implementation uses multiple channels in sequence: an automated WhatsApp message 30 minutes after abandonment for customers who have opted in, followed by an email 2 hours after abandonment for customers without WhatsApp opt-in, followed by a second email 24 hours after abandonment. That sequence, run via a connected AI system, reaches the maximum recoverable pool across the available channels. Running only one channel leaves recoverable revenue unclaimed.

How do you implement AI for abandoned cart recovery?

The implementation steps for AI cart recovery are: connect the AI tool to the store to receive real-time cart events, configure the follow-up message flows for each channel, train the chatbot on the product catalogue and the common objections for each product category, test the flows against real session data, and run for 30 days before drawing conclusions. The total setup time is 8 to 15 hours depending on catalogue complexity and how many channels you are activating simultaneously.

For Shopify stores, most AI cart recovery tools install via the app store and connect to the Shopify cart API automatically. Configuration of the message flows and chatbot knowledge is the manual work. For WooCommerce, the integration is typically via a plugin that connects to the WooCommerce order and cart events. The message flow configuration is similar.

Frequently asked questions

Does AI cart recovery work for stores without WhatsApp opt-in?

AI cart recovery via email still outperforms manual email sequences because the personalisation layer produces higher click-through and conversion rates than generic messages. A well-configured AI system sending personalised email about the specific products left, at the right time, with the right offer, recovers 7 to 9% versus 5 to 8% for a standard sequence. The WhatsApp channel adds the larger lift, but email personalisation alone is a meaningful improvement over the out-of-the-box sequence.

How quickly does AI cart recovery produce results?

A properly configured AI cart recovery setup produces measurable results within the first 30 days. The first week is typically lower performance as the system calibrates, and the second and third weeks show the rate settling toward the long-run number. Stores that evaluate recovery tools in the first 7 days before the system has run enough sessions are measuring noise. 30 days is the minimum evaluation window.

What does AI cart recovery cost?

AI cart recovery tools for ecommerce typically cost between 50 and 300 per month depending on message volume and the channels supported. WhatsApp Business API costs additional per-message fees (approximately 0.04 to 0.06 per conversation in most markets). The ROI calculation: if a 150 per month tool recovers 5 extra carts per month at a 60 average order value, the first month of recovered revenue is 300. The payback period is one month at that modest recovery count. At 50 extra carts recovered per month, the tool generates 3,000 in additional revenue for 150 in subscription cost.