Organizations have spent the past few years experimenting with AI. The question now isn't whether AI can generate insights. It's whether organizations can consistently turn those insights into business decisions that create measurable value.

That's where many AI initiatives stall. Models perform well in development but never make it to production. Teams generate insights but struggle to operationalize them across the business. And without governance, even successful AI projects can become difficult to trust or scale.

At SAS Innovate 2026, Product Manager Diana Maris explored how decision intelligence helps close that gap in her session, "How SAS^® Intelligent Decisioning Powers the AI-Driven Enterprise."

Here are four lessons from the presentation.

Lesson 1: AI doesn’t fail because of models, but due to the lack of strategy and operationalization

AI is just the latest technology trend that organizations are following, after trends like cloud
computing, social media and the blockchain. Each innovation has real value, but often
comes with hype where companies attempt to implement without a clear strategy. Like how many companies adopted cloud solutions, the same leadership teams are buying fragmented tools with no cohesive path to production.

There is data on this, too: A 2025 Gartner study found that only 41% of AI models ever reached production. That means lost opportunity and missed ROI, which are costly in a fast-changing business world.

To succeed with AI implementation, follow what successful vendors in the cloud era did: start small, prove the concept and then scale. Only then will you be able to gauge progress and then decide how to move forward.

Lesson 2: Decisioning forces alignment across silos

Today’s enterprise organizations can operate in silos, with teams passing like ships in the night. While the metaphor has become a cliché, it also hurts how businesses can operate efficiently.

Decisioning breaks down those barriers by forcing different teams to bring their data, models and business rules into one place to work towards outcomes that benefit the organization. Teams can no longer collaborate independently, resulting in an organizational and technical shift that helps everyone move forward together.

Lesson 3: AI without governance isn’t scalable or responsible

Many organizations are wide-eyed about AI because of its automation capabilities. However, we believe differently: AI helps enhance work, not replace it.

That’s because decisions impact real people, from home loan approvals to health treatment plans to manufacturing equipment. Decisions never sit in a silo, nor should they be outsourced to an agent; rather, they should be accountable and explainable. They should also be governed, which ensures traceability and control over automated decisions.

SAS’ decision intelligence capabilities provide governance capabilities, including audit trails and explainability, so anyone from a data scientist to a non-technical staff member can explain the outcome of a decision. The alternative is a human-out-of-the-loop, where no one can trace back what happened.

Lesson 4: Insights don’t drive ROI, actions do

I’ve visited numerous data and AI conference booths over the years, and a common theme that keeps showing up is the dashboard. Some vendors will tout a beautiful webpage of data, while others promise their platform will help generate better outcomes, with the actual action left up to the user.

With decision intelligence, insights can be translated into ROI and at scale. Decisions can also be reused, providing the scale needed to meet the business's needs across industries.

Decisioning is the missing AI layer

AI is still very new, so getting the strategy right is doubly important. Many organizations are searching for the right steps forward, and decision-making is that missing layer that connects insights and ROI.

Organizations that operationalize decisions to get the full value of their data and AI investments. To learn more:

Watch the SAS Innovate session