Representing Prompting Patterns with PDL: Compliance Agent Case Study
This work addresses the problem of inflexible and complex prompt engineering for programmers, offering a declarative solution that improves productivity, though it appears incremental as it builds on existing agentic programming concepts.
The paper tackled the complexity of prompt engineering for LLMs by introducing the Prompt Declaration Language (PDL), a novel approach to prompt representation that enables manual and automatic tuning, and demonstrated its utility in a compliance agent case study with up to 4x performance improvement.
Prompt engineering for LLMs remains complex, with existing frameworks either hiding complexity behind restrictive APIs or providing inflexible canned patterns that resist customization -- making sophisticated agentic programming challenging. We present the Prompt Declaration Language (PDL), a novel approach to prompt representation that tackles this fundamental complexity by bringing prompts to the forefront, enabling manual and automatic prompt tuning while capturing the composition of LLM calls together with rule-based code and external tools. By abstracting away the plumbing for such compositions, PDL aims at improving programmer productivity while providing a declarative representation that is amenable to optimization. This paper demonstrates PDL's utility through a real-world case study of a compliance agent. Tuning the prompting pattern of this agent yielded up to 4x performance improvement compared to using a canned agent and prompt pattern.