Preprint
Jul 2026
FORCE-Bench: A Benchmark, Dataset, and Evaluation Harness for Agentic AI in Enterprise Finance
This work introduces FORCE-Bench, which contains 251 expert-annotated queries and evaluates responses using a rubric-based framework calibrated to the requirements of the operational finance domain, across eight dimensions: accuracy, citations, clarity, depth, groundedness, recency, relevance, and structure.
Wolfgang M. Pauli, Sarah Panda, Kidus Admassu et al.
· 2 citations