Industry

Equestrian Sports & Digital Entertainment

Technology Used

PHP 8.3 / Magento 2.4.x
Google Gemini API
MySQL
Cron-Driven Lifecycle Automation

Project Solution

Design and delivery of an AI-powered retention system that tracks user behavior from the moment of registration, generates personalized email content through an AI engine, and draws on a video-derived horse personality knowledge base so every message feels specific to the user's own horse.

The Client

Our client is a mission-driven organization in the equestrian sports industry, dedicated to rescuing, rehabbing, and rehoming retired racehorses. Their digital venture built a first-of-its-kind gamified platform where users can own fractional shares in real racehorses at no cost, removing the financial risk and paperwork that traditionally kept everyday fans out of horse ownership.

Unfortunately, their platform's growth model was stalling. Registration and horse selection worked well, but engagement fell quickly after day one, and a great concept alone was not translating into a lasting community.

Platform Integration Challenges

There was an extreme drop-off between initial sign-up and sustained engagement, with a large share of newly registered users never returning after their first session. All users received the same generic outreach regardless of which horse they owned or which features they had used, resulting in low open rates and no personalization.

The platform also had no behavioral feedback loop to detect early disengagement, and horse personality data was static rather than dynamically surfaced, leaving the platform's core emotional hook underused. Any outreach required manual effort, a bottleneck that could not scale with user growth.

The Solution

This venture required a system that works in the background from the moment a user registers, tracking behavior, generating personalized content, and learning from real horse footage automatically.

Technologies Implemented: A combination of PHP 8.3 on Magento 2.4.x, Google Gemini AI, and MySQL were utilized in this project.

The Process: In order to fix a disengaging, one-size-fits-all outreach model, Chetu developers:

Built a cron-driven notification engine that triggers personalized emails at Day 1, Day 3, and Day 7, then weekly through an 8-week retention cadence.

Compared each user's activity delta against their overall profile to detect drift and generate re-engagement content before churn set in.

Ingested facility and horse behavior video through a Gemini-powered pipeline, converting footage into a structured, dynamic horse personality knowledge base.

Connected every platform action streaming, voting, badges into a single behavioral profile that feeds directly into what the AI engine says and when.

The Impact

The platform now operates far more personally and proactively. Day-7 retention improved by 38% post-launch, while the share of users still active at Week 4 increased from approximately 22% to 54%.

AI-personalized emails achieved a 3.2x higher open rate than generic outreach, along with a 47% increase in the adoption of previously unused features.

Manual outreach was eliminated entirely. Sixty-two percent of users remained active through the full eight-week cadence. Every new horse and new user enters the system automatically, with no additional staff effort required.

The same retention architecture can support any digital platform where users develop an ongoing connection with a specific entity, whether that is a horse, team, athlete, character, fitness goal, or digital experience.