Ep. 98 - Decisions, Decisions ... with Tony Phillips

There is a thing that I have not seen managed well in today's organizations: business rules. Typically, they are an “afterthought” in implementations (because you have to feed the business rules engine and control the process flow).But how many organizations rationalize the underlying business rules on a larger scale? When was the last time the magic “above $50 and below $50” rule was looked at? Remember $50 in 1979 is $222 today. And how do we check business rules and compliance when we don't even know what AI agents even do in detail?Today we are speaking with Tony Phillips, an experienced consultant, business architect, and business systems analyst with a focus on improving business efficiency and effectiveness. He covers strategy planning, capability modeling, process reengineering, systems analysis and design, data modeling, business rules engineering, and project management.Tony's experience, skills, and versatility allow him to excel as an architect, analyst, designer, mentor, trainer, and methodology implementer. He has a track record of providing critical services to and maintaining excellent relations with a wide range of clients, including government and financial services companies worldwide.In this episode of the podcast, we talk about:Defining Business Rules: Tony defines a business rule as a simple “if/then” statement that determines an output based on an input. He argues that business rules are a critical part of the “why” column in the Zachman framework and that they change more frequently than processes.Rules in Action: The hosts and guest discuss how business rules often change, using examples from the insurance and mortgage industries where the process remains the same, but the rules for approval constantly evolve.Separating Rules from Process: The discussion highlights the importance of managing business rules separately from processes and data. Tony notes that rules can apply to various aspects, including data types, data entities, and the relationships between them.DMN and Process Flow: The conversation touches on the Decision Model and Notation (DMN), a standard that connects an external rules table to a decision task within a BPMN diagram. This approach helps avoid complex, unwieldy decision trees.The Challenge of Complexity: The hosts and Tony discuss the complexity of modeling decisions, noting that a single, massive spreadsheet is typically not the most effective approach. Instead, they suggest using smaller, more focused decision tables that feed into one another.Governance and Maintenance: The podcast raises the critical question of who governs and manages business rules. Ideally, rules should be stored in a central, reusable repository where changes can be made and updated everywhere.False Positives vs. False Negatives: Tony explains the difference between false positives (a rule approves something it shouldn't have) and false negatives (a rule rejects something that should have been approved). He notes that false negatives are particularly difficult to spot and often require a manual audit or comparison to old methods.Business Rules and AI: The conversation transitions to the role of business rules in the age of AI. Tony proposes that business rules act as algorithmic “yes/no” decisions, whereas AI is more goal-oriented and based on probabilities.AI's Guardrails: The hosts and Tony agree that business rules are essential “guardrails” for AI, ensuring that AI's outcome-oriented behavior doesn't lead to unintended negative consequences. They also emphasize the need for human oversight and governance of AI-driven processes.Tony can be found on LinkedIn here: https://www.linkedin.com/in/anthonynphillips/.Please reach out to us by either sending an email to [email protected] or signing up for our newsletter and getting informed when we publish new episodes here: https://www.whatsyourbaseline.com/subscribe/.