The article argues that traditional metrics for measuring AI product success, like growth and retention, still apply but require a fresh approach, especially concerning engagement. Typical SaaS metrics like monthly recurring revenue (MRR) and customer lifetime value (LTV) remain relevant, but the author emphasizes the need for more inventive engagement metrics. For instance, at Elise AI, tracking the number of times users say "Thank You" to the AI offers a meaningful insight into user satisfaction, which could be more indicative of success than standard thumbs-up/thumbs-down ratings. The author suggests various creative metrics for evaluating AI tools, particularly in contexts like legal contract drafting. Potential metrics include the Edit-to-Save ratio, which measures how many edits lead to a saved contract, and the Clause Survival Rate, tracking how many contract clauses remain unchanged after negotiation. Other ideas include measuring late-night drafting activity to gauge user engagement patterns and using AI to assess the quality of drafts. These alternative metrics aim to capture the value derived from AI interactions, moving beyond mere usage statistics to understanding user behavior and satisfaction.