The Curious Case of Planning for Unreliable Agents: Challenges and Opportunities in Orchestrating Generative AI Agents
Published in Workshop on Planning in the Era of LLMs at the International Conference on Automated Planning and Scheduling (ICAPS-26), 2026
Recommended citation: Daneshi, R., Patra, S., Dwarakanath, K., Gopalakrishnan, S., Borrajo, D., & Sreedharan, S. (2026). The curious case of planning for unreliable agents: Challenges and opportunities in orchestrating generative AI agents. In Workshop on Planning in the Era of LLMs at ICAPS-26. https://scholar.google.com/citations?view_op=view_citation&hl=en&user=So86Wl4AAAAJ&sortby=pubdate&citation_for_view=So86Wl4AAAAJ:GnPB-g6toBAC
Abstract
Generative AI agents powered by large language models exhibit immense versatility but suffer from non-deterministic failures, hallucinations, and uncalibrated uncertainty. This paper investigates orchestrating swarms of unreliable generative agents using symbolic automated planners, formalizing recovery strategies, execution monitoring, and dynamic replanning protocols to guarantee multi-agent workflow reliability.
Recommended citation: Daneshi, R., Patra, S., Dwarakanath, K., Gopalakrishnan, S., Borrajo, D., & Sreedharan, S. (2026). The curious case of planning for unreliable agents: Challenges and opportunities in orchestrating generative AI agents. In Workshop on Planning in the Era of LLMs at ICAPS-26.