Workflowops improves multi agent task workflows using learned collaboration patterns
WorkflowOps: Learning Agent Collaboration Priors for Multi-Agent Workflow Orchestration
Artificial Intelligence
Summary
Coordinating multiple agents to work together on complex tasks is hard, and usually each new task starts without remembering past successes. The authors introduce WorkflowOps, a system that learns how agents have successfully collaborated before to guide new task planning. It also automatically creates new specialized agents when needed and uses smart matching to assign tasks efficiently while reducing reliance on large language model calls. Their experiments show that WorkflowOps improves the success rate in complex, structured tasks by reusing past collaboration patterns.
What this means in practice
- •For enterprise automation teams: Design automated multi-agent workflows that recall past collaboration patterns for improved efficiency and success on complex business tasks.
- •For software orchestration engineers: Develop dynamic agent pools by generating specialized agents on demand, enabling flexible handling of new task capabilities in software pipelines.
Authors
Qi Cheng, Shengyu Chen, Wei Cheng, Zhengzhang Chen, Xiaowei Jia, Haoyu Wang, Haifeng Chen
Abstract
Multi-agent systems are increasingly deployed for complex knowledge work, yet their orchestration layers remain largely memoryless: each new task is decomposed, assigned, and executed from scratch with no benefit from prior successful executions. We present WorkflowOps, a multi-agent workflow orchestration framework that learns agent collaboration priors from historical workflows and expands its agent pool on demand to cover new capability requirements. Our approach introduces three coupled mechanisms. First, a transition probability matrix captures pairwise agent collaboration frequencies from past workflows and applies them as soft guidance during DAG workflow construction through intra-layer ordering optimization, probability-thresholded edge suggestion, and transitive reduction for parallelism maximization. Second, a sufficiency-driven agent creation loop detects capability gaps via semantic matching scores, generates specialized agents through an LLM, and simultaneously injects them into the collaboration matrix, so that newly created agents are immediately usable with predicted collaboration priors. Third, a layered semantic matching strategy uses pre-trained sentence embeddings for fast, deterministic capability matching as a first pass, invoking LLM verification only for low-confidence cases, thereby reducing LLM routing calls by over 80\% compared to pure-LLM approaches. Experiments on mixed code, math, and question-answering suites show that WorkflowOps improves end-to-end pass rates over recent workflow-construction baselines, with the largest gains on structured, decomposable tasks where past agent handoff patterns transfer.