July 2025 Summaries

In a rapidly evolving landscape of AI-driven customer support, adapting leadership and processes to effectively leverage AI metrics has become crucial. This shift is exemplified by the introduction of the CX Score, a tool designed to analyze conversations and provide a comprehensive view of support experiences. Unlike traditional metrics like CSAT, the CX Score handles a significantly larger data volume, necessitating new strategies for data analysis and interpretation, such as using it for trend reviews and targeted evaluations. A key challenge in this transition is accepting the "black box" nature of AI, where the focus shifts from understanding detailed processes to appreciating broader outcomes and insights. Embracing the emotional aspect of these changes is also essential, allowing teams to process feelings and celebrate achievements. As AI technology continues to advance, including improvements in explanations and summaries, the benefits of AI metrics are increasingly outweighing the drawbacks, encouraging experimentation and adaptation in customer service approaches.

Intercom emphasizes the importance of "heartbeat metrics" as a critical tool for assessing system reliability, especially given the complexity of modern platforms and the need for immediate customer impact detection. These metrics provide a direct indication of whether a product is fulfilling its core function, offering a clear and fast signal when customer experiences are affected, thus enabling swift action before customers even report issues. At Intercom, the central heartbeat metrics include the rate of new messages and replies, reflecting the system's ability to support essential communication functions. These metrics not only inform the company's service level agreements (SLAs) by defining what constitutes downtime but also guide automated responses to incidents, such as rolling back recent code changes when issues are detected. By tracking these vital signs, Intercom ensures that its systems remain aligned with customer needs, fostering trust and accountability.

In May 2025, Intercom's product design team held a hackathon in London, aiming to create an Intercom-related feature or product using AI. Designers from across Europe participated, forming teams to explore tools like Lovable and Cursor. One team, including Daria Vorontsova and Tanya Ermolaeva, decided to build a website for "Intercom Design" to share the team's work and insights. Initially using Lovable, they encountered challenges with non-draggable and non-scrollable content blocks, leading to a switch to Cursor, where they rebuilt the site using Vanilla JS for better customization and speed. Despite the initial setbacks, the team completed a minimum viable product (MVP) within hours and won the hackathon. Over the following weeks, they refined the prototype with features like interactive maps and dark mode while planning content strategy. The revamped Intercom Design hub is now live, showcasing the team's ability to quickly create a functional and engaging platform with minimal engineering assistance.