From UPOS to AI: How Retail Technology Standards Are Evolving

Retail standards are entering a new phase as cloud, mobile technologies and AI reshape how systems communicate. Fiscal Solutions’ webinar explored the evolution from UPOS and POSLog to digital receipt standards, retail ontology and AI-ready models designed to support the next generation of global retail.

Retail technology has changed dramatically over the past three decades, yet many standards supporting today’s POS systems, devices and transaction data were created long before cloud platforms, mobile devices and artificial intelligence became part of everyday retail. This gap was the central topic of Fiscal Solutions’ webinar, “Retail Technology Standards: From UPOS to the Future of AI-Driven Retail.”

Presented by Roland Zališevskij, co-owner and Board Member of Fiscal Solutions, with an introduction from CEO Darko Pavić, the session explored how retail standards developed, why they still matter and what AI could mean for the next generation of retail systems.

Why retail standards were created

More than 30 years ago, retailers and POS software developers faced a major integration problem: different hardware devices often required different interfaces. A solution connected to one printer, scanner or cash drawer could require new development to support a similar device from another manufacturer.

The Association for Retail Technology Standards (ARTS), originally operating within the National Retail Federation, began developing common standards for devices, data and terminology. The goal was to simplify integrations and provide a common foundation.

One of the first important standards was OPOS, created for devices such as printers, scanners and cash drawers in the Microsoft environment. JavaPOS later extended the concept to Java-based environments. These approaches eventually came together under UnifiedPOS, or UPOS, which remains in use today.

POSLog and digital receipt standards

Another major ARTS standard is POSLog, an XML-based model designed to define and exchange retail transaction data. It created a common structure for information generated by POS and cash register systems.

Over time, POSLog became very large. The webinar noted that the model includes more than 7,000 unique elements. This provides broad coverage, especially for large software providers and Tier 1 retailers, but can also make implementation complex.

The ARTS Digital Receipt standard addressed another need: representing receipt information in a structured electronic form. It can contain data that may be exchanged, stored, printed and reused in processes such as loyalty programs, coupons, warranties, returns and exchanges.

The existing standard is XML-based, but work is underway on a new JSON specification. The webinar stressed that this is not simply a format conversion. The aim is to redefine the digital receipt for cloud and mobile environments.

A structured digital receipt may serve not only as a customer-facing document but also as transaction data for other systems, including fiscal reporting to government authorities. This is particularly relevant in fiscalization, where transaction information must often be adapted to different country requirements.

From ARTS to OMG and retail ontology

The organization behind the standards has also evolved. ARTS standards moved to the Object Management Group (OMG), which has broader experience in developing and maintaining technology standards.

One of the most important current projects is retail ontology. In simple terms, ontology is a machine-readable vocabulary that defines retail concepts, their meaning and the relationships between them.

This addresses a practical international problem. Companies can encounter different words for the same fiscal or retail concept, or the same word used with different meanings. This creates confusion in legal documentation, software requirements, localization and cross-border projects.

A shared retail ontology can create more consistent terminology across systems and markets. Fiscal Solutions applies this principle in its Fiscal Requirements Portal, professional materials and consulting methodology, helping harmonize terminology across jurisdictions.

UPOS 2 for cloud and mobile retail

Another important development is UPOS 2, which aims to modernize the traditional device-integration model.

Older UPOS architecture was designed around devices connected directly to the same machine as the POS application. Modern retail is different: POS functionality can operate in the cloud or on smartphones and tablets, while several applications may need access to the same physical device.

The proposed UPOS 2 architecture therefore moves toward REST API-based communication. Devices could receive jobs through a network, queue requests and serve multiple POS systems rather than relying on a direct local connection.

The work is still ongoing, and the webinar highlighted a wider problem: technology is moving faster than traditional standardization processes.

AI creates the next challenge

Even while the industry is adapting standards to cloud and mobile technology, artificial intelligence is introducing another major shift.

AI-driven retail requires more than technical APIs. If AI agents are expected to understand business processes, make decisions and coordinate activities across applications and devices, they need a shared understanding of retail concepts and data.

This is where semantic models and retail ontology may become increasingly important. The webinar described a future in which common data models provide the language, APIs expose system and device capabilities, and AI provides intelligence and orchestration.

Model Context Protocol (MCP) was also mentioned as one possible mechanism for connecting AI with applications and APIs. Future standards may therefore need to support not only communication between software and hardware, but also AI agents capable of understanding context and coordinating retail processes.

Why standards still matter

The webinar’s main message was that standards are far from outdated. Growing technological complexity makes them increasingly important.

For retailers and technology providers operating across multiple markets, standards can reduce integration effort, improve terminology, support reusable data models and make expansion into new countries easier. Fiscal Solutions explained that using standardized models can also help software solutions support scenarios from markets where a company may not yet have direct project experience, because the required concepts can already be represented in the common structure.

At the same time, standardization itself must become faster and more adaptable. AI is developing particularly quickly, while standards are traditionally created through slower, step-by-step processes. The challenge will be to create common models that are stable enough to provide interoperability but flexible enough to keep pace with technology.

As retail moves from locally connected POS hardware toward distributed cloud systems and AI agents, the role of standards is expanding. What began as a way to connect printers and scanners is becoming a common language through which people, applications, devices and intelligent systems can understand and work with retail data.

Ema Stamenković, Portal Manager at Fiscal Solutions

Linkedin profile