Business Rules are First Class Objects with James Taylor, CEO of Decision Management Solutions
In this episode, host Jimmy A. Hewitt had the opportunity to chat with James Taylor, CEO of Decision Management Solutions
This episode, we have the dynamic pair, Jimmy A. Hewitt, and the brilliant James Taylor, CEO at Decision Management Solutions. 💼 James Taylor, a true leader in the business rules space, will guide us through the rationale behind this game-changing perspective of why business rules should be treated as first-class objects within your organization!
Guest: James Taylor, CEO of Decision Management Solutions
Background and Entry into Decision Management
James Taylor’s background includes working in business roles. He has 20 years in the Business Rules field
Recommended business rules to engineering teams in various software companies. He was frustrated how often engineers chose to write code rather than adopt business rules.
Joined a business rules management system company due to frustration.
Systems were developed for banks, combining business rules, predictive analytics, machine learning, optimization, etc.
Coined the term “decision management” to describe this technology stack.
Focus of the Company: Automating Complex Decisions
The goal is to automate decisions in scenarios where the best course of action isn’t immediately clear.
These decisions need to be automated effectively and in a manageable way.
Evolution of Decision Management
Initially, building these rule based systems was expensive, focusing on areas like credit card fraud.
Value of combining technologies became evident, leading to coining the term “decision management.”
Advances in technology, data, and the cloud reduced costs significantly.
Decision management is now accessible for lower-volume transactions and diverse industries.
Accessibility and Opportunities
Decision management is now more affordable and applicable to various contexts.
Opportunities are widespread; it’s about identifying the right use cases.
Many industries can leverage these capabilities, given the lowered cost barriers.
Could you provide an overview of the landscape of decision management and business rules, particularly in the context of intermediate to advanced topics like high-value applications and predictive decisioning using AI/ML?
Overview of Decision Management and Business Rules
Decision Management and Business Rules Focus: The main topic is decision management, particularly centered around business rules. While initially managed through a business rules management system (BRMS), the term “decisioning platform” is now used for better clarity.
Decisioning Platform Analogy: The analogy used likens the decisioning platform to a “Process Management” platform rather than just a “Task Management” one, underscoring the importance of decisions in business processes.
Breaking Down Complex Problems
Purpose of Decisioning Platform: The decisioning platform’s primary purpose is to break down complex problems into more manageable components. Examples include loan origination, claims processing, and customer offer decisions.
Simplification with Business Rules: Business rules, which involve logical expressions with conditions and actions, simplify the formulation of complex decisions into more understandable terms.
Structured Decision Making
Role of Business Rules: Business rules serve as building blocks, enabling the assembly of rules that guide decision outcomes. This structured approach empowers organizations to make complex decisions using a simplified rule-based framework.
Execution of Business Rules
Execution Process: Executing business rules is not the main challenge, as computers excel at logical operations. Various platforms can handle rule execution, shifting the focus towards rule management.
Importance of Rule Management
Rule Management Emphasis: The focus is on effective rule management beyond execution. This involves identifying rules for modification, ensuring safe changes, and understanding the potential impact of alterations. Cost-efficient implementation of changes is crucial.
Role of BRMS (Business Rules Management System)
Comprehensive Management Framework: BRMS serves as a comprehensive framework for managing rule-based decisions, offering tools to simplify rule management.
Simulation and Comparison: BRMS facilitates simulating the effects of rule changes, comparing different rule versions, version control, and granting specific access to stakeholders.
Simplified Management with BRMS
Streamlined Rule Management: BRMS provides tools and capabilities to streamline decision rule management. It allows simulation of rule changes’ effects, clear comparison of rule versions, versioning, and role-based access control. The goal is to simplify rule management, making updates and changes more straightforward.
How do organizations currently manage their existing business rules, which are often already being executed, and what challenges do they face in terms of handling these rules, improving their sophistication, making changes, understanding the impact of changes, and determining the current state of their rule management process?
Importance of Effective Rule Management
Emphasizes the necessity of rapid accessibility and modification of business rules for functionality.
Current Rule Management Approaches
Some organizations rely on hard-coded programming languages like Python and Java for rule execution.
A notable percentage of businesses still manage rules manually through policy documents, guidelines, and scripts.
Progression to Advanced Rule Management
Organizations face the challenge of advancing from manual rule management to more sophisticated methods, possibly involving coding.
Integration of Logic into Processes
Many organizations have fragmented logic dispersed throughout their processes, within rules management systems, process engines, or custom code.
Historical tendency to break complex decisions into numerous process steps, each involving specific rules and conditions.
Shift Towards Decision Modeling
The practice of breaking decisions into multiple process steps has diminished due to the introduction of decision modeling alongside process modeling.
Decision modeling has improved clarity and reduced complexity in process design.
Legacy Systems and Complex Processes
Legacy systems still exhibit the practice of fragmenting business decisions into multiple process steps, resulting in complex process flows.
Example from claims handling demonstrates the challenges posed by having multiple instances of invoking business rules within a process.
Challenges of Entwined Decision-Making
Managing processes with interconnected decision-making steps makes implementing changes complex and challenging.
What does it look like for decision for a company who does treat decisions as a first-class object?
Simplified Processes: By considering decisions as distinct steps, processes become simpler and more agile. Decision-centric design condenses multiple rule-focused process steps into a single decision step, streamlining the workflow.
Analytics Integration: Treating decisions as primary elements enables the seamless integration of machine learning and predictive analytics. The data-driven nature of analytics aligns well with a decision-focused approach.
Enhanced Agility: Decoupling decisions from processes significantly increases agility. Changes in decision logic can be made independently, without requiring modifications to the entire process.
Stability through Separation: Separating decisions from processes stabilizes workflows. Many processes perceived as highly variable have variable decisions within them. This separation simplifies change management.
Overcoming Legacy Challenges: Extracting decision logic from legacy systems helps manage complexities and aids platform replacement. This simplification facilitates transitioning to new platforms.
Overall, considering decisions as primary entities in business operations leads to streamlined processes, smoother analytics integration, enhanced agility, and greater stability—critical advantages in today’s dynamic business landscape.
How can you help companies transition from their existing decision logic embedded in platforms, like loan origination systems, to a more improved state, whether by harvesting existing logic or starting anew, and what does this transition involve?
Must Believe in Decision Modeling
Strong belief in decision modeling.
Early adopter of the decision model invitation standard.
Consistently using decision modeling since around 2011.
Advocates for decision models due to their speed, accuracy, and engagement.
Building Decision Models
Always build decision models in projects.
Focus on talking to the business about current decision-making processes.
Importance of understanding how decisions are made today.
Reverse engineer existing logic into decision model frameworks.
Not relying solely on code; business perspective is crucial.
Handling Existing Logic
Discussing existing decision logic embedded in systems.
Example of origination project with scorecards and auto reject rules.
Convert low-level rules into decision artifacts.
Establishing the right decision structure and details.
Incremental process of building the model.
Automating Current State
Aim to automate the current decision-making process.
Adds value by increasing consistency and embedding in software.
Goal: Higher rate of automation, reduced manual work.
Preparation for conversations about improving decisions.
Exploring Improvements
Transition to discussing improvements in decision-making.
Involvement of machine learning, prediction analytics, AI.
Focus on automating for enhanced decision accuracy.
Importance of Automation
Emphasis on automating decisions first.
Establishing an automation framework as a foundation.
Boosts confidence in decision-making processes.
Step one: Achieve automation and consistency.
Common Approach
Common approach involves direct interaction with business.
Build top-down decision models for manual processes.
For existing automation, combine top-down model with existing logic.
Discovery of inconsistencies during this process.
Example of Inconsistencies
Example of inconsistencies in credit risk decisions.
Variation in credit risk acceptability based on FICO score.
Highlighting the complexity of intertwined rules.
Need to address discrepancies and align decision-making.
Reuse and Consistency
Step Two: Transition to discussing improvements in decision-making.
Increased reuse and consistency through well-defined decisions.
Decisions as more reusable artifacts compared to fragmented logic.
Continuous Improvement Framework:
Step Three: Implementation always includes a continuous improvement framework.
Avoids the common issue of not empowering business to manage rules after initial implementation.
Vertical Slices Approach:
Projects involve building more vertical slices of functionality.
Workshops to refine rules and test them.
Create simulation environments to assess the impact of rule changes.
Develop dashboards for analyzing production decisions.
Data-Driven Decision Improvement:
Turn on production provides data about past decisions.
Data informs how to make better decisions in the future.
Business users can analyze data and make changes.
Case Study: Claim Team in Hong Kong:
Weekly review of claims.
Identify tweaks to reduce manual reviews.
Historical data simulation for proposed changes.
Present results to supervisors for approval.
Deploy changes regularly to improve decision-making.
End Goal
The end goal is an agile, data-informed decision-making process.
Improve decision handling and process efficiency on an ongoing basis.
How does AI/ML intersect with BRMS?
Human In the Loop vs. Human On the Loop
Explains the concept of “human in the loop” vs. “human on the loop” in decision-making.
BRMS allows having a “human on the loop” approach, where humans are involved in decision improvement rather than being directly in the loop of each transaction.
Transition towards a human-governed AI approach.
Clarifying “Human Governed AI”
Introduces the concept of “human governed AI.”
Critiques the idea of involving humans just to blame them.
Emphasizes the importance of explicit definitions of the roles of humans and machines in decision-making.
Augmentation and Decision Context
Highlights the growing interest in augmentation, combining AI and human roles to enhance automation.
Importance of defining the specific contributions of humans and machines in decision-making.
Suggests that focusing on augmentation requires clear understanding of each entity’s role.
Use of Machine Learning in Business Decisions
Differentiates between areas where machine learning can replace existing rules derived from data analysis and areas requiring human judgment.
Describes using machine learning to predict conditions from medical reports to streamline manual reviews.
Emphasizes that machine learning for business decisions is complex and must be considered in the context of the decision process.
Failure Rates in Machine Learning
Mentions the high failure rates in machine learning and AI projects, especially when not contextualized.
Points out that the success rate in practical applications is low due to challenges in aligning machine learning with decision contexts.
Adding Value Through Decision-Making
Highlights that machine learning only provides predictions and decisions require actions.
Stresses the need for clear decisions based on predictions to add value.
Expresses that the significance of decisions is evident when considered in the context of business processes.
Machine Learning vs. Decisioning Platforms
Advises that a decisioning platform is crucial for operationalizing machine learning at scale.
Suggests that successful integration of machine learning requires a suitable decisioning framework.
Poorly Defined Requests and Context
Criticizes vague data requests from business executives.
Advocates for contextualized questions and understanding the implications of interesting findings.
Highlights the complexity of creating predictive models without proper decision-making context.
Simpler Machine Learning Problem
Discusses an example where machine learning can be applied to predict specific conditions in documents.
Points out that decision context simplifies the machine learning problem.
Reiterates that machine learning’s applicability depends on the specific decision context.
Challenges of Machine Learning in Decision-Making
Repeats the notion that machine learning’s success relies on decision context.
Addresses the common issue of spending more on machine learning without proportional benefits.
Echoes the need to integrate machine learning with decision-making processes.
What would you like the audience to take away from from this episode?
James emphasized the importance of decision modeling and automation. He advised beginning with the current decision-making process and then gradually automating it for better efficiency. He discussed the collaboration between humans and machines and also mentioned the challenges of applying machine learning in business scenarios. He highlighted the considerable value in decision automation and encouraged a shift in mindset regarding its complexity and benefits.
Watch Episode 26 Now
Have a Topic Idea? Let Us Know
If you have a topic idea, contact us today. We want our podcast to answer your questions, and if you have something that you want to hear, we would love to know. In addition, you can ask questions and get in touch with our Salient Process team.
Why Salient
At Salient Process, we understand the importance of comprehensive digital transformation. We work closely with each client and use a North Star methodology to align your procedures with overall business goals. Our expert team works hard to understand your unique needs, making a customized plan for success. With our Digital Business Automation solutions, you can move closer to your digital transformation goals.
Subscribe now
👉Subscribe to Bots & Thoughts: The Digital Business Automation Podcast Here
