SAS Innovate: The real AI advantage? It’s not what you think

Ask most individuals what offers AI its edge and so they’ll possible level to hurry, automation or the aura of generative AI instruments.

However in line with specialists at SAS Innovate 2025, AI’s actual aggressive benefit isn’t the algorithm – it’s the power to make use of it responsibly to make trusted, quicker, higher choices.

As SAS approaches its fiftieth anniversary, Chief Know-how Officer Bryan Harris used the milestone to not look again however to put out a daring path ahead: Constructing smarter, trusted techniques.

Listed below are 5 key takeaways from SAS Innovate’s opening session with Harris – filled with real-world examples and quotes that challenged typical eager about AI’s worth.

1. AI’s superpower is choice intelligence

Through the years, we’ve seen waves of technological breakthroughs: First got here the PC and web, then the cloud, machine studying and deep studying. Now, it’s all about AI – in each taste conceivable: generative AI, agentic AI and quantum AI. The speedy tempo of innovation can really feel overwhelming, even to the businesses constructing it.

“This tempo of innovation can really feel intense and overwhelming, even for the businesses that create it,” Harris mentioned. “However each subset of AI is fueling one factor: choice intelligence.”

The actual worth of AI lies not in what it creates, however within the choices it permits.

You’re not in search of simply any choice. You want a choice that delivers outcomes that outperform. You want a choice that helps you compete to win out there. You want a choice benefit. Bryan Harris, SAS CTO

2. Generative AI gained’t repair damaged enterprise fashions

Generative AI is making waves in areas like customer support and textual content summarization, nevertheless it’s not a magic repair for all enterprise issues. Harris referred to as out the parable that merely including generative AI to a course of will resolve all of your points.

“It appears like all it’s good to do is sprinkle a little bit generative AI within the enterprise and all your issues will magically disappear,” he mentioned throughout his keynote.

Whereas generative AI has spectacular purposes, it additionally has critical limitations, particularly when not utilized correctly.

“You’ll be able to have , unbiased AI mannequin – however feed it the incorrect knowledge and you’ve got a biased end result,” Harris warned.

LLMs and immediate engineering alone are usually not sufficient for many enterprise use instances. You want an orchestration of LLMs, machine studying, APIs and extra to make sure accuracy, equity and governance.Bryan Harris

The truth is, a current examine revealed that giant language fashions are likely to suggest greater rates of interest and mortgage denials for Black candidates in comparison with white candidates, even when credit score scores are the identical. On common, Black candidates want credit score scores 120 factors greater to get permitted on the similar charge, in line with the examine.

Harris emphasised that such discrepancies are systemic throughout main LLMs. A easy resolution, like including extra emphasis to the immediate, doesn’t resolve the difficulty.

“That’s why LLMs and immediate engineering alone are usually not sufficient for many enterprise use instances,” he mentioned. “You want an orchestration of LLMs, machine studying, APIs and extra to make sure accuracy, equity and governance.”

This factors to the shift in business focus from generative AI to agentic AI – shifting past simply creating content material to creating AI-driven choices which can be truthful, correct and efficient.

3. From generative AI hype to agentic AI actuality

Whereas the tech world nonetheless loves generative AI, SAS is already deploying what’s subsequent: agentic AI – techniques that purpose, act and collaborate with or with out human enter.

And it’s shifting quick. The worldwide agentic AI market is anticipated to exceed $70 billion by 2030, with 52% of organizations including AI brokers to workflows this yr.

So, what does SAS imply by agentic AI?

“SAS thinks of it as a spectrum between ‘human out of the loop’ and ‘human within the loop,’” Harris mentioned.

SAS thinks of [agentic AI]as a spectrum between ‘human out of the loop’ and ‘human within the loop. Bryan Harris

“Backside line … we don’t simply provide you with an agent – we provide you with transparency and explainability.”

At one finish, brokers act autonomously, like denying fraudulent transactions in actual time. Then again, they flag irregular instances, present danger scores and provide context for human evaluate.

That is shared intelligence in motion – already at work in SAS® Clever Decisioning. In a demo, Harris confirmed how SAS® Viya® can flag mortgage instances, clarify its reasoning and provides reviewers full transparency from mannequin playing cards to choice lineage.

These brokers aren’t restricted to fraud or loans. SAS clients use them throughout industries. They’re inbuilt low-code environments and mix LLMs, guidelines, fashions and APIs into scalable, reusable workflows.

The outcome? Smarter, safer and extra accountable AI with governance and equity inbuilt.

4. What if… digital twins may assume like operators?

What should you may ask your manufacturing unit flooring, “What if?” and get solutions backed by actual knowledge, simulation and AI?

That’s what SAS, Epic Video games and Georgia-Pacific are constructing collectively.

At Georgia-Pacific’s Savannah River Mill, which is over 400 meters lengthy (bigger than 4 soccer fields), they constructed a digital twin to check how autonomous guided automobiles (AGVs) transfer by means of the ability. These machines are vital, however discovering the best fleet measurement and routing is hard.

“As we speak’s manufacturing vegetation are advanced environments with AGVs, robotics and other people all working collectively,” Roshan Shah, VP of AI at Georgia-Pacific, defined. “When issues go incorrect, it will get costly and other people can get damage.”

Utilizing SAS Viya, the staff ran simulations inside Unreal Engine to check AGV methods. They assumed extra AGVs would enhance productiveness, however the digital twin confirmed in any other case.

“Including AGVs appears intuitive, proper? Extra employees ought to imply extra output,” Shah mentioned. “However in Unreal, the price of additional AGVs really slowed issues down.”

The optimum quantity turned out to be 47, boosting efficiency by 8%.

“As an alternative of constructing and testing these options in the true world,” mentioned Invoice Clifford, VP of Unreal Engine Ecosystem at Epic Video games, “one click on lets us discover one of the best configuration amongst all attainable worlds.”

SAS even added a “manufacturing unit editor” to the simulation, letting managers regulate layouts, reroute AGVs, or toggle between actual and artificial knowledge in a gamelike interface.

As Shah put it, “Think about being a plant supervisor and having the ability to regulate operations on this method. It’s an entire game-changer. Pun supposed!”

This isn’t simply visualization. That is choice making powered by AI and immersive tech.

Understanding digital twin know-how

5. Remodeling optimization with quantum AI

Harris pivoted to quantum AI as a game-changer for tackling large-scale optimization issues.

“Think about having a digital twin that solutions your most troublesome ‘what if’ questions immediately,” he mentioned, noting how quantum accelerates knowledge processing from hours or days to minutes.

He referenced longtime collaborations with international organizations like Procter & Gamble. “We’ve labored collectively for many years to enhance efficiency and cut back price,” Harris mentioned. “Now we’re exploring how quantum AI can take that to the following degree.”

As Krista Comstock, Director of Digital Innovation at P&G, put it, “One of the crucial important takeaways is the advance in productiveness and efficiency. Quantum, being a probabilistic drawback solver, would not at all times give one of the best reply, regardless of being quick. Whereas conventional solvers would give greater high quality solutions, however take longer to resolve. The hybrid method offers us one of the best of each worlds: each pace and high quality of the answer.”

From decreasing compute occasions to boosting the effectivity of digital twins, hybrid-traditional quantum approaches are already displaying actual promise in driving transformation at scale.

The underside line

AI could also be evolving quick, however SAS helps clients evolve quicker, with instruments that make each choice extra clever, explainable and efficient.

Whether or not it is by means of digital twins, agentic AI or quantum AI innovation, one factor’s sure: The long run belongs to those that construct belief into each layer of their know-how.

Need extra from SAS Innovate? Learn these weblog posts