Dialog Insight Email Send-time Optimization

My last product at Dialog Insight was an AI-powered send-time optimization tool designed to improve email engagement by delivering messages to each recipient at their personalized optimal time. Clients using the product saw a 40% increase in engagement for communications using the feature.

The idea originated from repeated conversations with customers about how they chose when to send emails. Most described a largely random process, relying on intuition or slow trial-and-error over weeks. Even when a time performed well for certain segments, engagement often dropped for others. This revealed a critical insight: send-time is a personalization opportunity where a one-size-fits-all approach does not work.

After validating this insight through customer discussions and engagement data, I researched how competitors approached send-time optimization. Most focused on identifying a single optimal time for all recipients. Given Dialog Insight’s differentiation around personalized communications, I concluded that we needed a recipient-level optimization strategy rather than a global one.

I led the development of a system that analyzed historical engagement data to determine each recipient’s optimal send-time. Instead of sending all emails at once, the system distributed delivery over multiple hours so each recipient received the message at their predicted best time. Designing this solution required addressing challenges such as data sparsity and reliability, which led us to implement fallback mechanisms and configurable rules to ensure decisions remained robust even when data was limited.

Before a full launch, we released the feature to a small group of clients to validate accuracy, scalability, and real-world impact. Based on feedback and observed performance, we refined both the prediction logic and delivery mechanics before rolling it out more broadly.

Beyond performance gains, this project taught me valuable lessons about AI-driven product design—especially the importance of user trust. Many users were initially hesitant to allow an automated system to make decisions that affected key business metrics. To address this, we focused heavily on transparency, giving marketers visibility into how decisions were made and when communications would be delivered. This proved essential in driving adoption and confidence in the system.