Reproduce “SPAR: Experience-Driven Strategy Evolution in Multi-Session B2B Sales” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hangming Zhang, Yao Zhao, Haotian Kuang, C. L. Philip Chen, Tong Zhang, Tengyue Ma
SPAR addresses multi-session B2B sales by separating merchant-specific memory from a shared playbook of hypotheses. An LLM labels outcomes as strategy-driven, external, or uncertain; only strategy-driven outcomes update the playbook. In a 1688 A/B test, SPAR reports relative lifts of 15.8% in conversion and 35.8% in M1 high-intent rate versus Direct Prompting. The test evaluates the system, not attribution accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.