When AI is too verbose, full of repetition and says sorry for being an AI…

This case study addresses the challenge of AI applications experiencing high user acquisition but struggling with retention, alongside practical issues with AI tool usability. Andrew Chen identifies that many AI-powered apps generate initial excitement and growth due to novelty, but suffer from high churn rates as users quickly abandon these tools.

Chen's approach involves both tactical optimizations and strategic insights. On the tactical side, he demonstrates how to improve ChatGPT's usefulness through custom instructions that eliminate verbose, apologetic responses and add follow-up questions to extend conversations. He also shows integration techniques like adding ChatGPT to Apple Watch shortcuts for faster access. Strategically, he analyzes the broader AI app ecosystem, noting that most are simple API wrappers lacking differentiation or stickiness.

The key insight for product managers is that AI applications follow familiar growth patterns seen in previous technology waves (web 2.0, mobile). While current AI apps benefit from novelty-driven acquisition, long-term success will depend on traditional retention metrics like D1/D7/D30 and network effects. Apps that merely wrap existing AI APIs without creating sustainable user engagement or integrating into existing workflows will likely fail as the novelty wears off.

Product managers should focus on building AI features that integrate into users' existing workflows, create network effects, and provide genuine utility beyond the initial "wow factor." Success will ultimately be measured by retention rather than just growth.