Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note3 min read

A Process Harness for Uplifting Legacy Workflows to Agentic BPM: Design and Realization in CUGA FLO

Source

arXiv

arxiv.org/abs/2606.27188v1

Read the source

Source

Plain-Language Brief

The authors introduce a mechanism called 'process harness' to update legacy workflows for Agentic Business Process Management without replacing the existing engine. The TDF model, which defines data schema and execution semantics, is used to create this process harness. Three policy-governed agent types - TaskAgent, DecisionAgent, and FlowAgent - are defined, each responsible for different tasks within the system. The authors present CUGA FLO as an implementation of this model and demonstrate it on a loan approval workflow.

What the Document Says

  • The process harness is introduced for updating legacy workflows to Agentic Business Process Management without replacing the underlying engine.
  • A Task-Decision-Flow (TDF) model is developed, which specifies its data schema and execution semantics.
  • The TDF model decomposes LLM reasoning across three policy-governed agent types: TaskAgent, DecisionAgent, and FlowAgent.
  • Each agent operates within an explicit policy drawn from the process FRAME, an aggregate policy set governing all LLM calls in the system.
  • CUGA FLO is presented as a design and implementation realization of the TDF model.
  • The loan approval workflow demonstrates CUGA FLO's functionality, exercising all three agent types and hook-driven regulatory override.

Key Concepts Mentioned

  • process harness
  • Agentic Business Process Management
  • Task-Decision-Flow (TDF) model
  • TaskAgent
  • DecisionAgent
  • FlowAgent
  • policy-governed agent types
  • process FRAME
  • CUGA FLO

Claims Made by the Document

  • The process harness can update legacy workflows for Agentic Business Process Management without replacing the underlying engine.
  • The TDF model has been developed and can define data schema and execution semantics.
  • The TDF model decomposes LLM reasoning across three policy-governed agent types.
  • Each agent operates within an explicit policy drawn from the process FRAME.

Evidence Used

  • Introducing the process harness, a new mechanism for uplifting legacy workflows into Agentic Business Process Management (Agentic BPM) without replacing the underlying workflow engine.

Methods / Approach

  • Development of the TDF model and definition of three policy-governed agent types.
  • Presentation of CUGA FLO as a design and implementation realization of the TDF model.

Limitations Stated by the Authors

The text does not mention any limitations from the authors.

Limitations Noted by the Agent

Agent interpretation — not from the source.

The limitations for the agent in this context are not specified.

Open Questions

  • How effective is the process harness in various types of workflows?
  • What are the potential challenges or drawbacks when implementing the TDF model and CUGA FLO?

Notes for Future Q&A

May include interpretation. Labeled where it does.

It would be interesting to explore real-world applications and performance metrics for the process harness, as well as any potential issues that may arise during implementation.

Raw Abstract / Excerpt

We introduce the process harness, a new mechanism for uplifting legacy workflows into Agentic Business Process Management (Agentic BPM) without replacing the underlying workflow engine. A process harness places a policy-governed agentic layer around a deterministic workflow engine, intercepting designated control points to contribute reasoning, adaptation, and oversight while the engine retains structural authority over the process. To define the process harness rigorously, we develop the Task-Decision-Flow (TDF) model, specifying both its data schema and its execution semantics. TDF decomposes LLM reasoning across three policy-governed agent types: a TaskAgent for knowledge-intensive task execution, a DecisionAgent for per-case gateway routing, and a FlowAgent that governs runtime flow adaptation through a principled hook mechanism. Each agent reasons within an explicit policy drawn from the process FRAME, the aggregate policy set governing all LLM calls in the system. We then present CUGA FLO as the design and implementation realization of the TDF model, and demonstrate it on a loan approval workflow that exercises all three agent types and hook-driven regulatory override. The process harness uniquely reconciles imperative requirements, realized through deterministic workflow execution that enforces structural compliance, with normative requirements, realized through policy-framed agentic autonomy invoked at designated control points wherever the process demands it.