SPARCs with SAS: Part 2–Model Development. Tools, Techniques, and the Citizen Data Scientist

Originally published on ARCweb.com by Colin Masson of ARC Advisory Group on January 16th 2025.

Welcome to the second in our three-part SPARC (Short Podcasts by ARC) series with SAS, where we'll be unpacking the complexities of adopting Artificial Intelligence (AI) in the industrial sector. As an industry analyst at ARC Advisory Group, I've had the opportunity to speak with numerous organizations, and one thing is clear: while the potential of AI is immense, so are the challenges. In this series, I’m exploring Industrial AI challenges with Bryan Saunders, Global Director and Head of IoT Industry Consulting at SAS to share some key insights.?

In SPARCs with SAS: Part 1—Data Management. Navigating "Any Data Anywhere" for AI Success, Bryan and I explored the critical need for a pragmatic approach to AI implementation, steering clear of the silver bullet mentality. In Part 2, we're delving into why Gen AI is not a universal panacea and why selecting narrow AI techniques is often more appropriate, and how to develop a practical AI roadmap by focusing on specific business outcomes.

This discussion is a must-listen for anyone serious about driving real-world value from AI in their organization.

Watch or listen to Part 2.

Tune into our SPARCs Channel on YouTube for Parts 1 and 3, and future SPARCs.

SPARC SUMMARY

Key Discussion Points

Bryan and I covered a range of topics, but several key themes emerged that are crucial for industrial companies:

• The Right Tool for the Job: Colin and Bryan agree that Generative AI is just one tool in the industrial AI toolbox. They both caution against the assumption that it's a solution for every problem, noting that proven "narrow AI" techniques are often more suitable, cost-effective, and explainable.
• Focused Use Case Approach: Bryan highlighted the importance of tackling one business outcome at a time. He shared examples of how SAS helps customers like Georgia Pacific and Lockheed Martin build out their AI capabilities by focusing on specific pain points and then expanding from there. This approach is particularly effective because it allows for incremental progress, faster results, and the development of a solid foundation for future, more complex projects.
• Data Dictates the Approach: The discussion underscored how the nature of your data influences the analytic methods you can apply. This is a critical consideration that is often overlooked.
• Empowering Subject Matter Experts: A key point from Bryan is the need to expand the resource pool by empowering subject matter experts and "citizen data scientists" with user-friendly tools to build and optimize models.
• Interpretability Over Accuracy: Brian made a compelling point that in industrial settings, model interpretability often trumps model accuracy. It's critical to understand why an AI model is making a certain recommendation, and that is not always the case with black box, unexplainable AI.
• The SAS AI and Analytics Lifecycle: We discussed the SAS AI and analytics lifecycle, and its strong alignment with ARC’s industrial AI use case lifecycle.

Why You Should Listen

This podcast isn't just another theoretical discussion. It offers concrete, actionable advice for industrial companies looking to leverage AI. You’ll gain:

• A clear understanding of why Gen AI is not a universal solution and when narrow AI techniques are more appropriate.
• Insights into how to develop a practical AI roadmap by focusing on specific business outcomes.
• An appreciation for the importance of data quality and how it drives the selection of analytic methods.
• Guidance on how to empower your subject matter experts in the AI process.
• Real-world case studies from industry leaders that showcase successful AI deployments.

Looking Ahead

In the concluding Part 3 of this series with SAS, Bryan and Colin explore the challenges and opportunities awaiting Industrial AI practitioners as they work to transform data and AI models into tangible real-world results.

Learn More About ARC and SAS Views on the AI and Analytics Lifecycle

• Download the ARC View titled Industrial-Grade AI: Transforming Data into Insights and Outcomes
• Follow Colin Masson and Bryan Saunders on LinkedIn for more insights and a transcript of today's conversation.

Interest SPARCed?

Thanks for tuning in to this SPARC! Share these episodes with others who are passionate about applying technology the right way to address skills gaps, create more intelligent processes, while building a more profitable?and?sustainable future. For those looking to go deeper, ARC Advisory Group offers longer format?Digital Transformation and Sustainability Podcasts,?and unparalleled guidance in developing digital transformation and sustainability strategies for industrial organizations. Reach out to us for expert insights and support that can help you turn these ideas into action and lead your organization toward a more profitable and sustainable path.

To contribute to SPARCs, or the Digital Transformation and Sustainability Podcasts, contact Colin Masson or Jim Frazer at ARC Advisory Group through your client manager.

For ARC Advisory Group recommendations for?closing the digital divide by embracing Industrial AI, and governing and guiding major decisions about enterprise, cloud, industrial edge and AI software, please contact?Colin Masson?at?[email protected].