From "Crutch" To Coach: Patterns of Sustained Engagement and Deepening Support in AI Wellbeing Coaching

Why It Matters

Sustained AI coaching demonstrates measurable emotional depth, suggesting scalable mental‑health support that moves beyond short‑term fixes and could reshape digital wellbeing services.

Key Takeaways

• 70.8% sessions revisit prior issues
• Regular users double deep engagement rates
• Work-related topics show highest continuity
• Complex issues stay single-session
• EXP ratings show moderate reliability (ICC 0.62‑0.79)

Pulse Analysis

The rapid adoption of on‑demand AI chatbots for mental health has sparked debate over their role: quick information crutch or long‑term coach. Nova’s recent analysis provides evidence that AI can transition into a coaching function, delivering continuity that mirrors human‑led therapeutic relationships. By tracking session‑to‑session continuity and employing the Experiencing Scale (EXP) calibrated against expert raters, the study quantifies not just usage frequency but the depth of emotional engagement, a metric traditionally reserved for face‑to‑face interventions.

Methodologically, the research leveraged large language models to classify topic continuity and generate EXP scores, achieving moderate to good inter‑rater reliability (ICC 0.62‑0.79). The data reveal that 70.8% of returning sessions revisit earlier concerns, and users with five or more interactions are more than twice as likely to reach deep engagement levels (EXP ≥ 5). Notably, wellbeing and work‑related topics exhibit the highest revisit rates, while more complex psychological issues are appropriately limited to single sessions, underscoring an emerging ethical boundary for AI‑driven coaching.

For industry stakeholders, these findings signal a viable pathway to scale personalized mental‑health support without sacrificing relationship quality. Companies can design AI platforms that encourage repeat interactions, integrate engagement metrics like EXP, and tailor content to topics that benefit from iterative coaching. Regulators and clinicians may also view the demonstrated reliability as a foundation for establishing standards around AI‑mediated wellbeing services, paving the way for broader acceptance and integration into corporate wellness programs.