PageInsider™
Development and rebuild of the PageInsider™ interface — an AI tool in Landingi that used a machine learning model to predict which elements of a landing page visitors would pay attention to. The project covered modernising the interface, developing analytics features and integrating with the A/B testing module.
– enlargeWhere it started
I joined the PageInsider™ project in 2021 while the feature was still in beta. Page analysis was handled by a machine learning model, but the existing React interface needed usability improvements and the code needed to be brought into order.
As the tool evolved, it also needed broader analysis capabilities, improved result presentation and integration with other landing page optimisation features.
The challenge
I was responsible for developing and rebuilding the PageInsider™ frontend, covering UX/UI improvements, state management reorganisation, new feature implementation and automated test development. I also had to integrate the tool with the Optimization module and A/B tests, then migrate it to the new SPA.
How I built it
I started by improving the interface and message copy, then rebuilt the module in React, moving shared state to the Context API. I rewrote the side panel presenting the Clarity Score, guide and recommendations, and the top bar for switching between the main page and the thank-you page. I also developed the preview carousel, Smart Map layer management and analysis state handling — including result generation, errors and incomplete screenshots.
I implemented interactive area of interest (AOI) selection, letting users check the predicted attention level for specific landing page elements. I added keyboard support, the ability to jump to the editor and compare results across analyses, a tour for new users and Whitelabel account support.
On the Optimization tab I built the PageInsider™ panel integrated with A/B tests and a comparison view with AI Variant™ including a predicted conversion rate. I also developed automated tests in Jest, Enzyme and Testing Library.
In 2023 I integrated PageInsider™ with the landing pages list and migrated the feature to the new React and TypeScript SPA. In 2024, after the decision to retire the tool, I removed it from the application and cleaned up the related code.
The outcome
The result was a comprehensive interface letting users analyse predicted visitor attention before publishing a landing page. With Smart Map, Clarity Score and interactive area selection, users could see which page elements drew attention, make changes in the editor and compare results across analyses.
Integration with the Optimization module brought attention analysis together with the other landing page optimisation tools. The feature was later migrated to the new SPA and retired from the product in 2024 by company decision.



