What are Decisions?

Having clear-cut success metrics where capital investment needs to be considered against other priorities helps justify future business cases and demonstrates value to build momentum.

what should you automate?

Measurable and Aligned

Although not always possible, wherever practical, an automation initiative with objectively measurable business outcomes and strong subjective metrics are preferred.

Having clear-cut success metrics where capital investment needs to be considered against other priorities helps justify future business cases and demonstrates value to build momentum. Also, and more critical, you want first to understand what the business priorities are, overall, what are the current strategic themes? Will the selected business case help to achieve the stated strategic objectives?

In business, there are always many areas that can be improved. Prioritizing, in terms of strategic impact, is important. E.g., if your business is switching from wholesale to consumer direct, you may not want to be investing to make the wholesale model more efficient.

The stated criteria are not specific to decision automation but important for all business improvement initiatives and often neglected in choosing a starting point.

To summarize, prioritize decision automation initiatives that:

  • Align with strategic objectives

  • Are objectively measurable

The following are specific criteria for decision automation project selection.

Please note: Not all of these need to be true.

Quick Criteria

A quick smell test can help to identify good candidates areas.

  • Enough transaction volume?

  • Many SME’s working on tasks that makes it hard to scale up quickly?

  • Business frustrated with the pace of change?

Enough operational transaction volume?

Although each transaction’s dollar value may be small for repetitive transactions, high volume transactions can add up to be of dominant financial importance in business. These are usually an easy target for decision automation.

Inversely complex strategic or tactical decisions that are seldom made don’t make for good automation candidates in an operational sense.

Decision support systems can better address these, Subject matter expert knowledge combined with ad hoc insights from analytics or machine learning.

High variability of interpretation

Complex or expert decisions made by subject matter experts can often result in different outcomes based on who is making the decision. These areas tend not to scale up well as it takes time to train up staff.

Automation of decisions will allow for better scalability and more consistent decision making.

Areas of Change

Complex or expert decisions made by subject matter experts can often result in different outcomes based on who is making the decision. These areas tend not to scale up well as it takes time to train up staff.

Automation of decisions will allow for better scalability and more consistent decision making.

Where does the automation of decisions fit?

Decisions vs. Process vs. Data vs. Machine Learning

There are so many standards and techniques for automation out there. BPM’s BPMN, Decision modeling with DMN. Data modeling techniques and a myriad of machine learning models. Process, data & machine learning predictions prepare an organization for decision making. To make smarter, efficient decisions, decisions should be modeled centralized and automated, resulting in more streamlined and simplified processes. We, therefore, start modeling decisions before processes in automation requirements gathering. We model decisions before selecting data, before defining processes, and before training machine learning models.

  • Processes exist to get data ready to make decisions.
  • Machine learning models make predictions that inform decisions but don’t make decisions in of themselves.
  • Data is there to support decision making.

Building a business case

Once you have decided on the most appropriate candidate decision for automation, most organizations require justification to secure funding or be considered for prioritization for funding. For-profit organizations do things typically for three reasons; Increase revenue/profit, reduce risk, and increase operational efficiency.

You would want to show enough movement on at least one of these three metrics. Take the time to identify which of these metrics would be improved by automating the decision in question and determine the best way to quantify the dollar value the improvement would bring.

Risk can be converted into a dollar value by looking at putative penalties or losses incurred based on the adverse outcomes. Whether you hire an external consulting company or if you have an internal capability already, project costs should then be calculated and contrasted against the expected gains as your return on investment or ROI.

Consider the example: Claims Automation

Costs

Manual assessment is costing a company 2 million a year in salaries and it’s estimated 400k in fraud and 600k in customers lost due to satisfaction issues (slow or inconsistent payments).

The company would also like to expand its offerings, but it’s expensive to scale up because all the hiring and training would need to be done before a new book of business is live. So that’s another $500k.

A total of 3.5 million a year in current costs.

 

Return

If we were to automated 70% of the system’s simpler claims and left 30% to the current assessors, it’s estimated that it will free up 50% of their time. Some of that can be put toward growth initiatives.

500k reduction in growth investment + 1m in cost savings + estimated 500k in fraud and customer retention for a total of 2m per year.

That’s 6m over 3 years.

 

Investment

Let’s assume you don’t have a decision automation capability yet and you are going to bring in a external consultancy for the first few projects.

  • Services costs $600k
  • Software $500k
  • Internal IT Costs $300k
  • Annual ongoing: $320k
  • Year 1: $1.4m
  • Year 2,3,n: $320k

 

Return on your Investment

3 year investment is $2.36m return estimated at $6m

Subjective: there are of course benefits outside of the above hard figures, such as better consistency, being able to scale-up or scale-down faster.

Freeing up valuable SME’s to focus on growth.

 

Every decision management or modeling project is unique.

The functionality offered by any DMN tool cannot be measured in a binary fashion. Nor can it be represented simply as a checklist of items which are part of the tool repertoire. It cannot be described on a single, graduated path either, because there are many independent aspects of functionality that may be more applicable in some business scenarios than others.

 

Follow the link below and experience how the DMN can be used within your unique business context, to assess and assist you in developing better business opportunities.

EXPERIENCE OUR INTERACTIVE USAGE SCENARIOS

Selecting a Decision Management System

It all begins with an idea. Maybe you want to launch a business. Maybe you want to turn a hobby into something more. Or maybe you have a creative project to share with the world.

Whatever it is, the way you tell your story online can make all the difference.

Base Features

Features that apply to most implementations and most organizations’ needs

Enterprise Features

Features that to enterprises that have larger teams, outsized execution requirements, or multiple teams that need to collaborate.

Advanced Features

Features that would support special requirements or advanced use cases where an organization has some capability in decision automation but wants to expand it’s competitive automation advantage

base features

  • Support for DMN standards including import/export. – Allows for skills interchange as well as maintaining central decision assets in a standard format.
  • Non-programmer decision authoring – allows for faster change and upskilling.
  • Decision table support – This is the most central artifact in business rules authoring, robust support for tables is foundational
  • Decision Lifecycle management – Ability to version, branch and manage decisions for releases
  • Deployment pipeline management – Functionality that facilitates and automates deployments of decisions to be consumed as a service or API
  • Scenario-based testing against defined expected results – this is similar to unit testing for code but tests logic in decisions against a set of expected results.
  • Tracing of execution – giving the ability to inspect after execution what decisions were made

enterprise features

  • Visual version comparison tools – Tools that allow for rapid change management of decision artifacts from the requirement to implementation components
  • Low Code object models – Easily update data models used by decisions without major code updates & deployments
  • Advanced/Large Decision table support – Ability to edit large and complex decision tables
  • Decision Catalog – Ability to manage shared decision assets across the enterprise
  • Full DMN model – Able to represent decisions in a model with shared decisions across multiple decision services

 

  • Allow traceability to original source – support for hierarchical knowledge sources that map decisions to original sources of logic e.g. regulations, machine learning models, etc. This enables fast impact analysis.
  • Requirements lifecycle management – ability to manage requirements lifecycle separately to the implementation components lifecycle, this facilitates collaboration across stakeholders.

 

  • Simulation – Able to run champion challenger simulations by varying the deployed decisions and running representative data sets against the decisions at scale.
  • Analyst-driven testing – Ease of use for business-driven testing rather than programmatic test cases
  • Containerized API-based deployment model – allow for the deployment of individual decision APIs
  • Hybrid cloud-ready – Support for deployment on all major cloud and virtualization platforms and ease of relocation from one platform to another

advanced features

  • Explainer XAI Support – ability to receive explainable artificial intelligence data and incorporate that into the decision-making explanations
  • Connect to models – ability to easily connect to versions of model APIs that have been deployed to a serving environment
  • Access feature farms – ability to access streaming grid data that allows for up to data variables to be used with decisions
  • Grid ready – Ability to deploy natively to grid compute like spark and have a licensing model that supports bringing the execution to the data.