INCOSE Orlando hosts Link Parikh to talk on AI in defense programs

By Cassidy Mills, Staff Writer

September 16, 2025

 

The Orlando Chapter of the International Council on Systems Engineering (INCOSE) gathered for a presentation titled “Advances in AI Applications: How the DoD Should Bring AI into Programs and Enterprises,” Sept. 10.

 

The presentation was delivered by Link Parikh, founder and chief executive officer of Rocket Technology Inc. It drew a diverse audience of engineers, program managers and industry professionals attending both in person and online.

 

The evening began with introductions from chapter members and guests, highlighting expertise ranging from defense contractors and government program managers to systems engineers, academics and recent graduates.

 

Parikh opened his remarks by warning against both the underuse and the overhype of artificial intelligence (AI). He said neglecting AI risks obsolescence while overpromising can lead to billion-dollar failures.

 

“AI is not a technology that sits on its own. It’s part of the mission,” Parikh said. “We’ve already seen cases where programs spent big on AI and collapsed because the integration was never feasible.”

 

He outlined two adoption models: augmenting existing systems with AI (“plus AI”), and designing new programs where AI is built in from the ground up (“AI plus”). He said the distinction is crucial for planning realistic investments and outcomes.

 

“When you try to force AI into everything, you end up creating more confusion than clarity,” Parikh said. “AI works best when it is tied directly to mission outcomes, not when it is added just to check a box.”

 

A recurring theme in his remarks was the rise of “shadow AI,” where employees use tools such as ChatGPT outside approved workflows. He said the practice, often driven by convenience and productivity pressures, creates risks by exposing sensitive data to public platforms and bypassing security controls.

 

“If you want to stop shadow AI, you have to provide a better product,” Parikh said. “Right now, people go to open systems because it works — and in defense, that means data is leaking into environments where it shouldn’t.”

 

To make sense of the shifting AI landscape, Parikh broke the technology down into three categories most relevant for defense: generative AI, machine learning and deep learning. He explained that each carries different strengths and risks, and that leaders need to understand the differences before aligning them with mission objectives.

 

“Generative AI is great for decision support,” he said. “Machine learning is about teaching a system to recognize patterns. Deep learning goes a step further — it can create things humans wouldn’t even think of.”

 

By drawing this distinction, Parikh urged the audience to see AI not as a monolithic capability but as a spectrum of tools to be applied with precision. He said treating AI as a diverse toolkit helps avoid hype and ensures resources are directed toward applications with measurable outcomes.

 

He then connected the technologies back to systems engineering fundamentals. He stressed that without the right frameworks, AI cannot be integrated in a sustainable way.

 

Parikh emphasized model-based systems engineering and modular open systems approaches as critical enablers for responsible adoption across large defense programs. These practices, he said, allow systems to evolve alongside emerging AI capabilities without requiring complete redesigns.

 

“Concurrency between development and simulation is essential,” Parikh said. “You cannot afford to separate those two anymore, because the moment you do, you fall behind. Simulation gives you the ability to test and refine ideas in near real time.”

 

His message underscored that AI integration requires a new tempo of engineering practice, one that closes the gap between design, testing and fielding.

 

Parikh reinforced the idea that AI’s true value lies in improving decision-making at every level of defense operations. From the desk of an analyst to the commander in the field, to the logic embedded in an autonomous system, he said AI should consistently aim to help humans and machines make better, faster and more informed choices.

 

“At the core, what we hope to always have is this — everyone making the most optimized decision at the desk, in the field and in the system,” Parikh said. “It’s not about replacing human judgment, but about amplifying it with insights that come from patterns no human could spot on their own.”

 

In closing, Parikh urged the audience to see AI not as a futuristic buzzword but as a present necessity. He said it must be deliberately engineered into systems with measurable returns to maintain superiority in complex, data-driven environments.