Kaizen and AI: Intercom Reinvents Customer Support with Fin Flywheel

Kaizen and AI: A Partnership for Optimized Customer Support

Imagine a world where each day brings an improvement, however small. This vision, though modest at first glance, can lead to significant transformations over the long term. This is the very essence of Kaizen, a Japanese philosophy of continuous improvement. Born in post-war Japan and popularized by companies such as Toyota, Kaizen is based on the idea that small, regular changes can lead to substantial results.

Kaizen does not merely promote incremental improvements; it is a structured approach to work, based on three fundamental principles. First, small changes are less intimidating and reduce resistance to change. Second, improvement must be a continuous process, not a one-time event. Finally, everyone in the organization must be involved, as those closest to the work are often best positioned to improve it.

In practice, Kaizen follows a simple yet effective cycle: identify a minor problem, test a change, measure the result, refine, and repeat. The goal is not to achieve revolutions but to make consistent progress.

The Fin Flywheel Model at Intercom

At Intercom, the Kaizen philosophy is applied to the AI agent, Fin, through a process called the Fin Flywheel. This model of continuous improvement unfolds in four steps:

• Train: Fin is trained to handle complex customer queries.
• Test: Simulated customer conversations are conducted to evaluate Fin's performance before deployment.
• Deploy: Fin is launched across all channels to ensure consistent customer support.
• Analyze: Fin's performance is continuously reviewed using AI-powered insights to enhance the customer experience.

This cycle is not a one-time process but a continuous loop where each interaction contributes to Fin's improvement. Rather than assuming that the AI will function perfectly from the start, improvement is built into the system itself. The more Fin is used, the better it becomes.

Integrating Human Insights

Continuous improvement at Intercom is not limited to AI. The Human Support team plays a crucial role in identifying issues that require human intervention. Whenever a customer problem is resolved, the team strives to understand why it could not be resolved by Fin and how to prevent its recurrence.

Support representatives are encouraged to suggest improvements, whether they involve content updates or process corrections. A lightweight system, supported by AI, allows for the quick recording and implementation of these suggestions, transforming the support team into a proactive improvement engine.

When a conversation reaches a human, essential questions are asked: why did this problem come to us? Why couldn't Fin resolve it? How can we prevent this from happening again?

A Sustainable Competitive Advantage

By integrating continuous improvement into daily operations, Intercom demonstrates that sustainable progress does not result from occasional transformations but from constant refinement. The Fin Flywheel and the involvement of human teams create a system where AI and people continuously learn from interactions with customers, thus providing a significant competitive advantage.

Kaizen shows that sustainable progress does not come from an occasional transformation; it arises from intentional and daily refinement. The Fin Flywheel applies this philosophy to AI. Our continuous improvement process for human support applies it to human insights. Together, they create a shared system where people and AI continuously learn from customer interactions.

When improvement is integrated into the mechanics of your work, it ceases to be a one-time project and becomes an ingrained capability. Over time, these small daily improvements not only add up—they lead to a competitive advantage.