Comparing the effectiveness of LLM-Powered coaching with human coaching and GPT conversation

Abstract

This study examines the effectiveness of LLM-powered coaching compared to traditional human coaching and GPT-4 interactions across 11 key criteria of coaching effectiveness. Leveraging a randomized controlled trial with 32 participants, the research revealed LLM-powered coaching significantly outperforms GPT-4 in empathy, engagement, and encouragement – achieving outcomes comparable to human coaching in all dimensions. Notably, LLM-powered coaching showed the greatest improvements in Positive Affect ( + 8.49%) and reductions in Negative Affect ( − 27.45%) among the groups. Qualitative feedback highlights the strengths of human coaching in emotional depth and nonverbal cues, while AI coaching excels in its accessibility, flexibility, and structured support for self-reflection and action planning. This research underscores the transformative potential of AI-powered coaching to foster individuals’ growth and well-being on a large scale. By offering accessible, scalable, cost-effective coaching solutions, AI coaching complements traditional human coaching, expanding access to personal and professional development for a broader and more diverse population.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 3 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.