Automatic message selection with a chatbot

What is this patent about?

’739 is related to the field of internet-based marketing and automated customer engagement. Specifically, it addresses the high rates of abandonment in e-commerce and lead-generation workflows, where users exit a website before completing a purchase or registration. The background context involves the limitations of static exit pop-ups and generic follow-up emails, which often fail to capture the specific reasons why a user decided to walk away from a transaction.

The underlying idea behind ’739 is to use an interactive chatbot as a diagnostic tool to capture qualitative feedback during the moment of abandonment and then feed that data into a remarketing engine. By engaging the user in a simulated human conversation at the point of exit, the system can extract specific objections—such as high shipping costs or lack of financing—and use these insights to select a highly targeted advertisement or incentive that directly addresses the user's stated barrier to entry.

The claims of ’739 focus on a method and system that triggers a customized chatbot within a messaging window when a user discontinues a transaction. The chatbot is dynamically tailored to the user’s demographic profile, including geographic location, age, or gender, and is designed to solicit a specific explanation for the abandonment. This captured explanation is then associated with the user's identity to programmatically select a relevant advertisement from a third-party database for delivery after a set period of time.

In practice, the invention functions by monitoring user behavior, such as cursor movement toward the browser’s close button or prolonged inactivity, to launch a chat interface that mimics a live agent. The system uses a neural network engine to parse natural language responses, allowing it to distinguish between different reasons for exit. For instance, if a user tells the chatbot that a product is too expensive, the system records this specific objection rather than just logging a generic failed conversion.

This approach differentiates itself from prior art by closing the loop between real-time conversational AI and downstream advertising. While traditional systems rely on broad browsing history or cookies to guess user intent, this invention utilizes explicit feedback gathered through dialogue. By integrating a management console that allows retailers to set specific campaign hierarchies and response delays, the system ensures that the subsequent remarketing effort is a direct, personalized counter-offer to the user's original reason for leaving.