Abstract
Conventional enterprise systems for creating business rules, for example sales compensation plans, can involve manual translation of strategic intent into structured configurations. This process may place a cognitive burden on users. The disclosed technology describes a system employing an orchestrated, multi-agent workflow to assist in this translation. A planning orchestrator agent can receive a user's intent in natural language, deconstruct the goal, and delegate sub-tasks to specialized agents. These agents may perform functions such as knowledge retrieval, parameter synthesis using retrieval-augmented generation, and compliance validation. The workflow can generate a validated, structured data object, such as a javascript object notation file, for execution by downstream enterprise systems. This approach to complex logic generation from abstract intent may help reduce manual configuration efforts, decrease errors, and improve alignment between business strategy and operational execution.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Ganaba, Marim; Gatlin, Jeannette; Smeltzer, Jarred; Moore, Dearbhla; and Pineda, Alex, "Agentic Workflow for Translating Natural Language Intent into Validated, Structured Business Rules", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/9219