Search papers, labs, and topics across Lattice.
This paper introduces FlowBank, a novel framework for optimizing agentic workflows in multi-agent systems by creating a reusable portfolio of complementary workflows instead of generating a single workflow per query. The authors demonstrate that their approach effectively balances performance and cost, achieving a 4.26% improvement over automated baselines and a 14.92% improvement over handcrafted ones across five benchmarks. By leveraging precomputed workflows and adaptive selection at inference time, FlowBank addresses the limitations of existing task-level and query-level methods, enabling more efficient and effective query handling.
FlowBank reveals that a compact, adaptive portfolio of workflows can outperform traditional single-query generation methods, enhancing efficiency and effectiveness in multi-agent systems.
Large Language Model (LLM)-based multi-agent systems are increasingly powerful, but current agentic workflow optimization paradigms make an unsatisfying trade-off. Task-level methods spend substantial offline compute yet deploy only a single workflow, leaving complementary candidates unused, while query-level methods synthesize a new workflow per query at substantial inference cost. Our motivating analysis shows these paradigms are more complementary than competing: workflows discovered during offline search often solve different subsets of queries, and many queries handled by expensive query-level generation can already be solved by cheaper precomputed workflows. This suggests a different objective: rather than searching for one universally best workflow or regenerating one per instance, we should build a compact bank of reusable, complementary workflows and select among them adaptively at inference time. Doing so requires solving three coupled problems: generating complementary rather than redundant candidates, compressing them into a small deployable portfolio, and assigning each query to the right workflow under a performance-cost trade-off. To this end, we present FlowBank, a three-stage framework for portfolio-based agentic workflow optimization. Diversifying proposes DiverseFlow to steer search toward under-covered queries and produce a high-coverage candidate pool. Curating proposes CuraFlow to compress this pool into a compact portfolio with minimal redundancy. Matching casts deployment as edge-value prediction on a query-workflow bipartite graph and routes each incoming query to the portfolio member with the best predicted utility. Across five benchmarks, FlowBank achieves the highest average score among the evaluated methods while remaining cost-competitive, improving over the strongest automated and handcrafted baselines by 4.26% and 14.92% relative, respectively.
