If you think of SAS as a data, AI and analytics powerhouse, Epic Games as the studio behind Fortnite and Georgia-Pacific as the company that makes paper towels and more, you’re not wrong.
But you’re also missing the bigger story. One that – quietly and collaboratively – is reshaping how industrial companies use real-time AI to solve real-world problems.
At the heart of this transformation is an innovative combination of AI, real-time analytics and 3D visualization to modernize manufacturing. Executives from SAS, Epic Games and Georgia-Pacific took the stage at SAS Innovate 2025 in Orlando to pull back the curtain on what they've been building – and why it matters.
The technology stack reads like a wish list for forward-thinking operations leaders: SAS® Viya® for advanced analytics and machine learning, Epic Games' Unreal Engine for hyper-realistic 3D visualization, and a connected digital twin infrastructure for Georgia-Pacific that mirrors real factory environments. But the magic isn’t just in the tools – it’s in how they’re integrated.
Keep reading about this collaboration or watch it in action in the video above. Or, both.
Digital twins, reimagined
We’ve heard the term “digital twin” tossed around in boardrooms and white papers. But as SAS Chief Technology Officer Bryan Harris pointed out on stage, it’s time to go deeper.
“Many people have different definitions. But the question we should be asking is: what is the goal of a digital twin for a business?”
For SAS, the answer is clear: digital twins should simulate complex environments, generate synthetic data and deliver actionable insights. Harris argued that digital twins should not only work like the real world but also look like it.
That’s where Epic Games comes in.
Learn more about digital twins
The Unreal advantage
Most people know Epic Games for Fortnite, one of the world’s most played games. They might not realize that the Unreal Engine, the technology behind Fortnite, powers everything from architectural renderings to autonomous vehicle simulations.
Bill Clifford, Vice President of Unreal Engine Ecosystem at Epic Games, explained it this way:
“Unreal is a real-time 3D engine used for product design, simulation, training, and sales and marketing applications across large enterprises.”
And now, thanks to a collaboration with SAS, Unreal Engine is powering photorealistic digital twins of real-world factories. These simulations aren’t just aesthetically pleasing – they’re smart.
The organizations co-developed a plug-in that connects Unreal Engine to SAS Viya. This allows users to access real-time data from simulated environments to fuel AI models that analyze, optimize and test industrial operations.
Testing the limits at Georgia-Pacific
Roshan Shah, Vice President of AI at Georgia-Pacific (GP), laid out the challenge clearly:
“Today’s manufacturing plants are complex environments with autonomous guided vehicles (AGVs), robotics and people all working together. When things go wrong, it gets expensive and people can get hurt.”
Their first test case was GP’s 400-meter-long Savannah River Mill – larger than four football fields. At this facility, which manufactures consumer goods including napkins, paper towels and toilet tissue, GP worked with SAS to optimize AGV routing for cost and safety.
The team surfaced a few promising scenarios by analyzing fleet movement using SAS Viya and modeling routing strategies. But instead of gambling millions on testing those strategies in a live factory, they turned to the digital twin.
“Thanks to Unreal Engine,” Clifford explained, “we can comprehend the complexities of the physical world in a high-fidelity virtual environment and iterate in real time with the simulation.”
Additionally, behind the scenes, SAS created a Factory Editor for the plant managers. Shah spoke about the impact of this:
“This editor makes it easy for plant managers to develop different ‘what if’ scenarios. It’s a complete game-changer. Pun intended!”
Increasing the number of AGVs seemed logical in one simulation, but it created a traffic jam. Unreal Engine flagged the issue and allowed the team to experiment virtually. The result?
“To improve the mean mission time by eight percent, we need to decrease the fleet size by three AGVs. That means the right number of AGVs in our Savannah facility is 47.”
Think about that: a fully simulated factory, providing a precise operational recommendation without lifting a wrench in the real world.
Safety without risk
But the innovation didn’t stop at logistics.
“A worker dies every 90 minutes from a job-related injury in the U.S.,” Harris said. “Today, existing computer vision models are only improving at the rate of workplace injuries.”
That’s unacceptable. So, the team asked, "What if we could train these models better without putting anyone in harm’s way?”
With Unreal Engine, they can. Digital workers were inserted into realistic factory scenes, engaging in normal and unsafe behaviors. Shah explained:
“We can simulate employees engaging in dangerous behaviors without posing risks to them in the real world. All of the behaviors you see here are automatically labeled as ‘unsafe.’”
They even altered factors like camera angles, lighting and protective gear to generate rich synthetic data that can power more effective safety models.
“Hopefully, an employee won’t be doing a push-up on the conveyor belt, but you get the point,” Shah joked.
The bigger picture
What makes this collaboration compelling isn’t just the tech – it’s the intent.
This collaboration shows how tools traditionally reserved for entertainment can be repurposed for business. It shows how immersive environments, synthetic data and AI can work together to save money and lives.
“This is the future of manufacturing operations,” Harris said.
And he’s right. Because once you can simulate anything, test everything and optimize without risk, why wouldn’t you?
This isn’t science fiction. It’s a working solution used by a global manufacturer, powered by technologies you already know.
The world of AI is accelerating fast. But innovative collaboration like this shows that we can build something better than the real thing when the right minds come together, with purpose and precision.
Not because it replaces reality, but because it makes reality work better.
