Understandable AI 101: Adding Transparency and Explainability to Your Business Strategies

[Watch the recording]

Understandable Decisions and Understandable AI 101

As businesses automate more of their business strategies, they will need to be more mindful of understandability. First, the underlying decisions must be transparent in a manner that they are traceable. Second, the underlying decisions must not only be explainable to developers and data scientists, but also non-technical employees and external stakeholders. In this webinar, we will focus on what we call Understandable Decisions. Drawing from Understandable AI frameworks, we will cover what organizations need to consider when automating decisions, regardless whether AI is used in decision-making. We will then go through a live demonstration of SMARTS™ decision management platform and showcase our Understandable Decision functionalities including play-by-play, audit trails, and real-time reporting and alerts.

What is Explainable AI?

Explainable AI (XAI) refers to an AI system where it is possible for humans to explain why the AI arrived at a particular prediction or decision. XAI was a response to the “black box” or opaque nature of machine learning models where even domain experts struggle to understand the models. XAI involves various processes and techniques to make the models more transparent. Typically this involves creating a second model, an explainer model, which approximates the inner workings of the first model. Then an interface displays the results of the explainer model in a manner that an expert can comprehend. Common methods are SHapley Addictive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME). However, XAI has its limitations. First, the existing methods are, at the end of the day, approximations. Second, usually only data scientists and engineers can understand the results of these methods.

Understandable AI is a response to the limitations of XAI, focusing primarily on expanding who can understand these models. The goal is to make AI accessible, transparent, and understandable to all stakeholders, including executives, product and project managers, designers, analysts, and ultimately customers. View this webinar to learn more about XAI, Understandable AI, and Understandable Decisions.

Additional Resources on Understandable AI:

• Webinar: Balancing Predictive Power and Regulatory Constraints in Automated Decisions through Decision Trees
• News: Understandable Decisions with Vienna, the new version of SMARTS™
• Blog Post: Generative AI Use Cases for Business Analysts

Best-in-class Series

Sparkling Logic’s best-in-class series webinars are designed for business analysts and decision management practitioners who may or may not be familiar with Sparkling Logic’s SMARTS™ business rules and decision management platform. We provide practical guidance through out the decision management lifecycle. Whether you’re looking for tips on how to design, test, deploy, monitor, or improve risk and other operational decisions, this webinar series is for you. You can view all our past webinars and sign up for our periodic newsletter to receive updates on upcoming webinars.

If you’re new to decision management and SMARTS™, here are some resources to get your started:

• First, read our business rules FAQs, decision management FAQs, and our Decision Management 101 blog series
• Next, watch our “Getting Started with a New Decision” video (and other videos) on our Product Demos page
• Then, register for our next live Product Tour