Will Large Language Models Become the New Operating System for Self-Service Kiosks? | AI Agent Trends in 2026

The rapid advancement of generative AI and AI Agents has sparked a new discussion across the self-service kiosk industry: could large language models (LLMs) eventually become the new "operating system" for intelligent terminals?

 

From a technical perspective, the answer is more nuanced than the question suggests. Today's AI is not replacing operating systems such as Windows, Linux, or Android. Instead, it is transforming the application and interaction layers by changing how users communicate with self-service devices and how business workflows are orchestrated.

 

For industries that rely heavily on self-service technology, this shift could redefine the user experience over the coming years.

 

 

For more than two decades, self-service kiosks have followed a largely consistent interaction model. Whether deployed in finance, healthcare, hospitality, retail, transportation, or government services, users typically complete tasks by navigating graphical menus, selecting options, verifying identity, making payments, and printing receipts.

 

This interface has served the industry well because it provides predictable workflows and standardized business processes.

 

However, recent advances in large language models have introduced a different interaction paradigm.

 

Instead of asking users to locate the correct menu, AI Agents can understand natural language requests such as:

 

"I'd like to check in."

"I need to schedule a medical appointment."

"Please print my receipt."

 

The AI interprets user intent before coordinating the appropriate backend systems to complete the requested task.

 

This evolution has led many industry professionals to ask whether AI could become the next software layer that users interact with first.

 

 

Despite the growing discussion, large language models are not replacing traditional operating systems.

 

Operating systems remain responsible for hardware management, device drivers, memory allocation, file systems, security, and resource scheduling. These foundational functions continue to be essential for every self-service terminal.

 

What is changing is the application layer.

 

Traditional kiosks guide users through predefined workflows. Every action—from identity verification to payment processing—is designed in advance.

 

With AI Agents, users can simply describe what they want to accomplish. The AI then interprets the request and coordinates the appropriate software modules while still following established business rules.

 

In other words, AI is becoming an intelligent workflow orchestrator rather than a replacement for the underlying operating system.

 

 

Unlike conventional PCs, self-service kiosks integrate numerous hardware components into a single platform, including cameras, microphones, ID readers, barcode scanners, payment terminals, printers, card dispensers, biometric authentication devices, and other peripherals.

 

Traditionally, these components have been controlled by individual software modules.

 

AI Agents introduce the possibility of providing a unified interaction layer capable of coordinating multiple hardware and software components based on user intent.

 

This approach is particularly valuable in public-facing environments where users have different levels of digital literacy.

 

For example, a Banking Self Service Kiosk could simplify account-related services by allowing customers to describe their needs instead of navigating multiple screens.

 

Similarly, a Healthcare Self Service Kiosk could assist patients by understanding appointment requests before interacting with hospital information systems.

 

In the hospitality sector, a Hospitality Self Service Kiosk may streamline check-in experiences by combining identity verification, payment, room assignment, and key card issuance into a more natural conversational workflow.

 

Government agencies are also exploring how a Public Service Kiosk can improve citizen services by reducing the complexity of administrative procedures while maintaining compliance with existing regulations.

 

 

Throughout the history of computing, interaction methods have continuously evolved—from command-line interfaces to graphical user interfaces, and now toward natural language interaction powered by AI.

 

This does not mean touchscreens will disappear.

 

Instead, natural language, voice interaction, and multimodal AI are expected to complement existing graphical interfaces rather than replace them entirely.

 

For many business scenarios, visual confirmation remains essential for displaying information, reviewing documents, confirming identities, and completing payment transactions.

 

The future is therefore likely to combine GUI and AI, allowing users to choose the most efficient interaction method for each situation.

 

 

Although AI Agents offer significant opportunities, several practical challenges remain.

 

First, regulated industries such as finance, healthcare, and public administration must continue to follow strict compliance procedures. AI can interpret user intent, but it cannot replace legally required business processes.

 

Second, successful deployment depends on seamless integration with existing enterprise systems such as PMS platforms, HIS platforms, ERP systems, and payment infrastructure.

 

Third, computing resources and deployment costs continue to influence implementation strategies. Depending on application requirements, organizations may choose cloud-based AI, edge AI, or hybrid architectures.

 

Finally, reliability and security remain critical considerations. Current industry practices generally position AI as the interface for understanding user intent, while established business systems continue to execute and validate critical operations.

 

 

Rather than replacing existing software architectures, AI is expected to become an additional intelligence layer.

 

Traditional architecture:

 

 

Future architecture:

 

 

Under this model, AI becomes the primary interaction interface while existing enterprise systems continue to manage business logic, compliance, and data processing.

 

 

Large language models are unlikely to replace traditional operating systems in self-service kiosks.

 

What they are beginning to change is how users interact with intelligent devices and how software components collaborate behind the scenes.

 

As AI capabilities continue to mature, the industry's competitive advantage may increasingly depend not only on hardware performance or software features, but also on the ability to deliver intuitive, efficient, and secure human–machine interactions.

 

For self-service kiosk manufacturers, system integrators, and solution providers, AI Agents represent not a replacement for today's technology stack, but an important step toward the next generation of intelligent self-service experiences.