A systematic literature review of 88 studies on LLM agents for production found that most research focuses on easily measurable aspects like task completion and accuracy, neglecting critical production-oriented factors. Key areas such as long-term reliability, failure recovery, monitoring, and resistance to attacks are reported as missing in over 88% of the studies. Even among studies with strong production evidence or those claiming deployment, these crucial operational aspects are frequently overlooked, indicating a significant gap between research claims and real-world readiness. AI
IMPACT Highlights a critical gap in LLM agent research, suggesting that current production claims may be overstated due to a lack of focus on essential operational metrics.
RANK_REASON The cluster reports on a systematic literature review published in a peer-reviewed journal, analyzing existing research on LLM agents. [lever_c_demoted from research: ic=1 ai=1.0]
- ACM Digital Library
- Docker
- IEEE Xplore
- Information and Software Technology
- Kitchenham–Charters
- LLM agents
- PRISMA-S: common questions on tracking records and the flow diagram
- Real-World Relevance Signals
- Scopus
- Web of Science
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →