General Atomics operates at the forefront of aerospace, defense, and advanced technology where precision, reliability, and scale are non-negotiable. Supporting that reality requires more than powerful systems; it requires applications that can evolve as quickly as the operations they serve.

That realization sparked a broader transformation.

In 2021, General Atomics adopted the Mendix platform to accelerate digital transformation and establish a long-term foundation for development. It started as an effort to modernize how applications were built and managed and became an opportunity to rethink how work gets done across the enterprise.

General Atomics has delivered more than 50 applications on Mendix, replatformed over 45 legacy systems, and built a growing community of 75+ certified developers. The organization has evolved its application environment into a governed, scalable platform for AI designed to help teams move from reactive processes toward more intelligent, agentic systems over time.

In doing so, General Atomics is positioning itself to close the gap between AI experimentation and real operational impact as its AI strategy continues to mature.

Evolving Enterprise Application Development

The path to intelligent systems began well before AI entered the picture.

In 2019, General Atomics’ application environment consisted of a mix of enterprise platforms, including ServiceNow, SAP, and internally developed tools. While these systems supported critical operations, they also introduced fragmentation with each solving a specific need without a unified approach to development, governance, or scalability.

By 2021, the need for a more cohesive strategy became clear. The team began evaluating low-code platforms to modernize development and create a more standardized, agile foundation.

The goals were clear:

• Rationalize and modernize existing applications
• Strengthen governance and establish a consistent operating model
• Embrace a cloud-first approach using microservices, prebuilt connectors, CI/CD pipelines, and AWS services
• Keep core systems clean while extending functionality through applications
• Shift from waterfall to agile delivery

At its core, the initiative aimed to accelerate digital transformation, increase application throughput, and meet stringent government cloud requirements.

To guide the decision, General Atomics leveraged Gartner Magic Quadrant and Forrester Wave evaluations.

Key factors in the final selection included:

• Native integration with SAP
• Alignment with manufacturing and industrial use cases
• Backing as part of the Siemens ecosystem
• Flexible deployment options, including GovCloud
• Built-in CI/CD automation

This combination made Mendix more than just a development tool, it became the foundation for enterprise-wide transformation.

Getting Started with Mendix

General Atomics’ first wave of Mendix applications focused on core operational needs:

• Field Support Request Application (FSRA): A centralized platform to request, review, and approve field support activities, improving visibility and coordination across teams.
• ITS Purchase Request (ITS PR): An application for submitting and tracking purchase requests, where administrators enrich request data and trigger integration with SAP to generate purchase requisitions.
• Surplus Inventory Marketplace (SIM): A system designed to track and manage the exchange of surplus inventory, creating greater transparency and efficiency in asset utilization.

While these applications addressed immediate use cases, they also served a broader purpose: establishing a consistent application development process.

With Mendix, General Atomics created a unified environment that provides real-time business context, maintains workflow state, and orchestrates processes across systems.

“The platform’s cloud-agnostic architecture allowed seamless integration with external services, including AWS and Azure, ensuring that applications, and future AI capabilities, could operate across the organization’s existing ecosystem without constraint,” said Peter Mendez, Mendix Solutions Architect at General Atomics.

This foundation would prove critical as the team began evolving toward more intelligent, agent-driven systems.

Laying the Groundwork for Intelligent Systems

As the application ecosystem expanded, even well-designed solutions began to show their limits.

Like many enterprises, General Atomics relied on a portfolio of applications supporting ticketing, hardware requests, and operational workflows. While effective individually, these systems remained largely reactive.

Users had to initiate actions, respond to alerts, and move between systems manually. Over time, this created friction.

Tasks slowed down. What should have been simple required multiple steps across disconnected tools. Context switching became constant, and interruptions reduced productivity.

The issue wasn’t capability, but cohesion.

At the same time, AI introduced new possibilities and exposed a deeper challenge. Early pilots showed promise, but remained disconnected from everyday operations, limiting their ability to scale.

Many systems still lacked the context, orchestration, and governance needed to turn intelligence into action.

For General Atomics, the challenge isn’t just adopting new technology. It is creating the foundation needed to eventually embed intelligence directly into operations so systems can move beyond supporting work to helping drive it safely and at scale.

A New Agentic Approach

Rather than layering AI onto existing systems, General Atomics rethought the foundation.

The goal was clear: progress from reactive workflows to proactive insights, and ultimately toward agentic systems capable of goal-driven execution.

Mendix became the enabling platform.

That context is critical. With Mendix, data is connected, contextualized, and orchestrated, ensuring AI operates as part of a continuous system, not in isolation.

Building a Modular, Controlled AI Ecosystem

At the core of this strategy is a planned modular, agent-oriented architecture.

Rather than relying on monolithic AI models, General Atomics is pursuing a composable approach using smaller, purpose-built agents.

“When you keep it focused, you get better results,” Mendez explained.

As these capabilities are developed, each agent will operate within a clearly defined framework:

• Context-aware – grounded in enterprise data and workflows
• Orchestrated – coordinated through Mendix-driven processes
• Modular – each agent performs a specific, well-defined task
• Human-guided – enabling collaboration between people and AI

This approach reduces ambiguity, improves accuracy, and supports incremental scaling.

It also reflects a broader shift toward hybrid workforces where people and agents operate together within governed, traceable systems.

Making Governance the Priority

From the outset, governance has been central.

General Atomics is designing clear guardrails for how future AI capabilities will operate, ensuring automation does not come at the expense of control. “This is really the most important piece,” Mendez emphasized.

Every proposed agent is intended to operate within a clear decision framework:

• Recommend – suggest an action
• Escalate – route decisions to the appropriate stakeholder
• Execute – perform an action within controlled limits
• Never do – explicitly restricted actions

Human validation remains essential, particularly for high-impact decisions.

With full traceability and auditability, the organization built a foundation of trust that’s critical for scaling AI in regulated environments.

Early Impact Across Operations

While the broader agentic vision is still evolving, the modernization foundation is already delivering value.

Process improvements from application modernization and workflow automation have reduced manual effort, helping teams spend less time navigating systems and more time focusing on higher-value work.

Just as importantly, General Atomics is establishing the governance, context, and architectural principles required for reliable AI outcomes as capabilities are introduced. The organization is evaluating and developing additional use cases, including alerts, email processing, and broader operational workflows.

The strategy remains consistent: Start small. Focus on high-value use cases. Keep agents well-defined. “Measure impact through time saved and cycle time reduction and scale deliberately,” Mendez noted.

From Workflows to Intelligent Systems

Today, General Atomics is laying the groundwork to move beyond traditional applications into a new category of enterprise systems where intelligence can be embedded directly into how work happens.

With Mendix at the core, General Atomics is creating a connected foundation where applications, data, and future AI capabilities can work together in a governed and coordinated manner.

Combining this foundation with a governance-first approach positions General Atomics to develop agentic systems that can move work forward safely, predictably, and within defined boundaries.