What is UPA?
UPA is an experimental tool for making an organization’s business rules, workflows, systems, evidence, assumptions, and unknowns explicit in a shared, machine-readable model. It is intended for people and AI agents working with organizational knowledge that is otherwise spread across code, documents, tickets, databases, and people’s heads. UPA can check selected model properties, keep rules traceable to evidence, and provide model context for AI-assisted work. It does not establish whether the model fully reflects reality.
UPA Features:
Business Rule Modeling: Define business rules and capabilities in reusable .upa specifications, then compose them into larger workflows.
Evidence Traceability: Connect modeled rules to their supporting evidence while recording assumptions, inferences, and unresolved gaps.
Formal Verification: Translate specifications into a canonical Logical IR and use Z3 to check declared properties under stated assumptions.
Change Impact: Explore how changes relate to a model’s connected rules, workflows, systems, and reusable components.
Model Viewer: Inspect a model through a hierarchy, narrative, or system map, with options to drill down into components.
Code Generation: Generate code for supported targets and provider-neutral interaction artifacts from verified models.
UPA Benefits:
Shared Context: People and AI agents can refer to the same explicit account of organizational behavior instead of piecing it together from disconnected sources.
Visible Uncertainty: Separating evidence, assumptions, and unknowns helps reviewers avoid treating an unconfirmed detail as established fact.
Earlier Consistency Checks: Formal checks can expose contradictions in modeled behavior before teams rely on the rules in later work.
Reviewable Decisions: Links between rules and evidence give domain experts a clearer basis for reviewing whether a model is credible.
Reusable Knowledge: Composing verified components lets teams carry established business behavior into larger workflows without rewriting each rule.
Use Cases:
Checkout Rules: Model order totals, basket behavior, or other checkout rules and check selected properties of the resulting specification.
Workflow Documentation: Turn policies and observed processes scattered across implementation and documentation into named, reviewable workflows.
AI Work Context: Give an AI agent a machine-readable account of relevant rules, evidence, assumptions, and open questions for a task.
System Change Review: Use connected components and rules to examine which modeled behavior may be affected when a workflow or system changes.
Model Exploration: Inspect the bundled eShopOnWeb storefront example as a macroprocess map, component hierarchy, or narrative.

