By Jennifer Kohlhepp, Managing Editor, ICMA 

Artificial intelligence (AI) is quickly becoming one of the most talked-about technologies in the global card industry, but at the ICMA EXPO AI Roundtable Discussion, one message came through clearly: while the opportunities are significant, successful adoption will require caution, structure and a strong understanding of the industry’s unique security demands. 

Across the discussion, participants explored how AI could improve everything from personalization and scheduling to maintenance, quality control, forecasting and training. At the same time, they emphasized that the secure card industry cannot afford to treat AI as just another off-the-shelf tool. In a sector built on trust, compliance and data protection, AI must be approached with both ambition and discipline. 

The Opportunity: Smarter Operations Across the Card Lifecycle 

Roundtable participants identified a wide range of practical and strategic use cases for AI in card manufacturing and personalization environments. 

One of the biggest areas of opportunity is in production intelligence. Manufacturers are already surrounded by valuable data from presses, personalization equipment and other production systems. AI offers the potential to turn that raw data into actionable insight—helping teams identify patterns in errors, predict machine failures, improve scheduling and reduce scrap. 

Participants pointed to predictive maintenance as a particularly promising application. Instead of reacting to equipment failures after they happen, companies could use AI to detect warning signs earlier, identify trends in parts wear and recommend maintenance before downtime occurs. In high-volume, time-sensitive production environments, that could translate into major efficiency gains and cost savings. 

AI also shows promise in quality control, where it could help detect recurring issues, flag likely problem areas on similar jobs and improve consistency across runs. For personalization operations, that could mean greater confidence that the right card is matched to the right individual and that data, credentials and visual elements are accurate and consistent. 

Another major theme was forecasting and resource optimization. AI could help companies better anticipate material needs, energy use, labor requirements, data storage demands and inventory levels. Rather than over-ordering supplies or holding more material than necessary, manufacturers may be able to use AI to make smarter decisions about what they need, when they need it and where resources can be reduced without increasing risk. 

The discussion also highlighted AI’s role in training and knowledge access. In many operations, onboarding skilled technicians and maintenance staff takes months or even years. AI tools could shorten that learning curve by providing quicker access to training materials, answering routine questions and helping standardize institutional knowledge. For companies struggling with workforce development, this could be a meaningful advantage. 

Beyond the factory floor, participants also noted opportunities in graphic design, market research, procurement and planning. AI can help synthesize large amounts of information, simplify complex data and support faster decision-making. For marketing and business teams, this means quicker access to industry insights and faster analysis of trends, products and customer needs. 

The Challenge: Security, Trust and Information Leakage 

While the roundtable surfaced real enthusiasm, it also revealed deep caution—especially around data security. 

Several participants emphasized that in secure card environments, public AI platforms are not trusted with sensitive information. Intellectual property, customer data and production details are simply too valuable and too sensitive to expose to open systems. As a result, many companies see the safest path forward as developing or deploying closed, internal AI systems rather than relying on public tools. 

This concern is especially pronounced in highly secure environments such as HSA-related operations and other tightly controlled manufacturing contexts. Even internally developed AI systems, participants noted, must be carefully contained and monitored. The industry is not yet ready to give AI unrestricted access to sensitive systems, and for good reason. 

The issue is not just cybersecurity. It is also about control, compliance and accountability. AI can generate recommendations, surface patterns and automate tasks, but companies still need people who understand the business, the machinery and the regulatory environment to validate those outputs. As one theme of the conversation made clear, expertise is not becoming less important in the age of AI—it may become even more important. 

The Reality Check: Walk Before You Run 

Another valuable takeaway from the roundtable was the recognition that not every company is ready for the most advanced AI use cases. 

While participants brainstormed ambitious possibilities, they also acknowledged that many organizations are still in the earlier stages of digital transformation. Before AI can deliver on its full potential, some companies may first need stronger data infrastructure, more connected systems and better internal processes. 

That is why the discussion repeatedly returned to a practical message: start simple. 

Rather than chasing the most futuristic applications first, companies may be better served by beginning with focused, lower-risk use cases—such as training support, forecasting assistance, maintenance alerts or quality trend analysis. These kinds of applications can provide immediate value while helping teams build confidence, governance practices and internal expertise. 

Participants also stressed the importance of executive sponsorship. AI adoption cannot be treated as an isolated experiment happening in one department. It needs leadership support, clear goals and cross-functional involvement, including compliance and risk teams. In an industry where mistakes can carry major consequences, oversight must be built into every stage of adoption. 

Human Judgment Still Matters 

If there was one consistent thread through the discussion, it was that AI should enhance human capability—not replace it blindly. 

Participants described AI as a tool that could act like a personal assistant, help users make more decisions faster and support more efficient operations. But they also recognized the limits of the technology. AI can help gather information, identify trends and generate suggestions, but it cannot independently guarantee accuracy, truth or appropriateness. 

That means human review remains essential. Whether the application is market research, maintenance planning, materials management or personalization quality assurance, experts are still needed to confirm that the data is sound and the conclusions are valid. 

In other words, AI may help companies work smarter, but it does not eliminate the need for skilled professionals who know how to ask the right questions and interpret the answers responsibly. 

A Defining Moment for the Industry 

The ICMA EXPO roundtable made one thing clear: the card industry is only at the beginning of its AI journey. 

There is already broad recognition that AI has the potential to improve efficiency, reduce costs, strengthen quality control and unlock new forms of intelligence across the card lifecycle. But there is equally strong recognition that adoption must be deliberate, secure and grounded in operational reality. 

For the global card industry, the future of AI will likely not be defined by hype alone. It will be defined by how well organizations balance innovation with security, automation with oversight and ambition with readiness. 

That balance may be the industry’s biggest challenge—but it is also its greatest opportunity.