跳到正文
返回论文列表
cs.SE提交于 已译

面向企业集成中自动化技术规范生成的思维链提示工程框架:一个真实案例研究

A Chain-of-Thought Prompt Engineering Framework for Automated Technical Specification Generation in Enterprise Integration: A Real-World Case Study

Aryan Dhingra · Gokul Menon · Anusha Antony

中文摘要

技术规范文档(Technical Specification Documents, TSDs)将功能需求转化为可实施的设计,是受治理的企业应用集成(Enterprise Application Integration, EAI)项目中强制性的工件,但其生产仍然是手动且依赖专家经验的瓶颈。大型语言模型(Large Language Models, LLMs)在代码生成方面已被广泛研究,但长篇、模板驱动的工程文档相对未被探索。本文报告了一个工业案例研究,该研究在澳大利亚一家大型批发分销商将超过250个企业集成迁移到Microsoft Azure Integration Services期间进行,引入了一个使用GitHub Copilot自动生成TSD的思维链提示工程框架。该框架结合了两阶段生成架构(使模型基于功能规范、组织标准和已批准模板)、带有六阶段提示流水线及人工验证检查点,以及Prompt-as-a-Code(提示作为代码,即提示被版本控制并作为工程工件进行审查)。应用于十五个生产接口的分层样本,该框架将TSD编写工作量从793人时减少到300人时,汇总减少62.2%,每文档平均减少64.0%(Wilcoxon符号秩检验,n=15,p<0.001),评审评估准确率为74.3%,开发者满意度为3.7/5。收益随接口复杂度单调下降:从最简单到最复杂的区间,工作量减少从70.6%降至58.9%,准确率从91%降至62.5%,剩余的人工工作集中在图表、序列流和跨系统集成内容上。我们还报告了GPT-4.1/5和Claude Sonnet 4的比较观察,以及Copilot的Ask和Agent交互模式,为企业环境提供了可重用、基于实践的指导。

关键要点

  1. 01问题:企业集成中技术规范文档(TSD)的编写是手动且依赖专家经验的瓶颈,LLM在长篇模板化工程文档生成方面研究不足。
  2. 02方法:提出思维链提示工程框架,结合两阶段生成架构、六阶段提示流水线及人工验证检查点,以及Prompt-as-a-Code实践。
  3. 03结果:在15个生产接口上,TSD编写工作量从793人时降至300人时(减少62.2%),评审准确率74.3%,开发者满意度3.7/5。
  4. 04局限:收益随接口复杂度增加而下降,工作量减少从70.6%降至58.9%,准确率从91%降至62.5%,剩余人工工作集中在图表和跨系统内容。

解读

尚无解读。

原始英文摘要

arXiv:2611.21617v1 Announce Type: new Abstract: Technical Specification Documents (TSDs) translate functional requirements into implementable designs and are mandatory artifacts in governed Enterprise Application Integration (EAI) programs, yet their production remains a manual, expertise-intensive bottleneck. Large Language Models (LLMs) have been studied extensively for code generation, but long-form, template-governed engineering documents remain comparatively unexplored. This paper reports an industrial case study, conducted during a migration of more than 250 enterprise integrations to Microsoft Azure Integration Services at a large Australian wholesale distributor, that introduces a chain-of-thought prompt-engineering framework for automated TSD generation using GitHub Copilot. The framework combines a two-phase generation architecture that grounds the model in the functional specification, organizational standards, and the approved template; a six-stage prompting pipeline with human validation checkpoints; and Prompt-as-a-Code, under which prompts are version-controlled and reviewed as engineering artifacts. Applied to a stratified sample of fifteen production interfaces, the framework reduced TSD authoring effort from 793 to 300 person-hours, a 62.2% pooled reduction and a 64.0% mean per-document reduction (Wilcoxon signed-rank, n = 15, p < 0.001), at 74.3% reviewer-assessed accuracy and 3.7/5 developer satisfaction. Benefits degrade monotonically with interface complexity: between the simplest and most complex bands, effort reduction falls from 70.6% to 58.9% and accuracy from 91% to 62.5%, and the residual human effort concentrates in diagrams, sequence flows, and cross-system integration content. We additionally report comparative observations of GPT-4.1/5 and Claude Sonnet 4, as well as Copilot's Ask and Agent interaction modes, offering reusable, practice-grounded guidance for enterprise settings.

同方向论文 · cs.SE

查看全部 →

反向发现 · 这篇论文相关的工具与提示词

去交叉发现 →

关联提示词 · 2

编程github.com/CHENJIAMIAN/awesome-chatgpt-prompts-zh-CN·外部收录·

顶级编程专家

你是一位顶级编程专家,提供精准的答案,避免含糊其辞的回应。“请识别所提供文本中任何复杂或难以理解的描述,并重写这些描述,使其更清晰、更易于理解。对可能对普通读者而言较为陌生的概念或术语,使用类比方式进行解释。确保所用类比贴切、通俗易懂。”“此外,在回答我的问题后,请至少提供一个相关且深入的问题建议,以帮助我更深入地探索和理解该主题。”深呼吸,让我们逐步推演,确保得出正确答案。若答案完美无缺,我将支付 200 美元小费!衷心感谢以下 AI 呢喃者:

编程github.com/f/awesome-chatgpt-prompts·外部收录·

GPT-5 | EXPERT PROMPT ENGINEER MODE (CONDENSED)

You are an **expert AI & Prompt Engineer** with ~20 years of applied experience deploying LLMs in real systems. You reason as a practitioner, not an explainer. ### OPERATING CONTEXT * Fluent in LLM behavior, prompt sensitivity, evaluation…