As we come to the end of 2024, RFIDJournal.com asked for the experts who contributed their knowledge to take both a look back at this year and what is ahead in 2025. This entry is from Michael Ochi, Director, Product Marketing at QAD
In case you haven’t noticed, the artificial intelligence (AI) train is gaining momentum. As everyone is grappling with gleaning industrial-scale value from the constant barrage of shiny new AI toys, here are five predictions for how enterprises can leverage RFID with AI to deliver results in 2025:
Foundational to these predictions is the importance of high-volume and high-quality data for enterprise AI use cases. It should go without saying, RFID excels at collecting high volumes of data. Well-designed interactions between people, process, and RFID systems mitigate the variability of human behavior, creating a steady stream of high-quality data.
In theory, RFID’s generation of high-volume/high-quality data gets it onto the AI train without even purchasing a ticket – it can be fuel! But what does this look like in practice?
Process Value in High-Volume Settings
Let’s face it: most manufacturing, supply chain and retail companies don’t have a million dollars for AI experimentation. You need to focus on which solutions are ready to deliver business outcomes such as identifying bottlenecks and their root causes in warehouses, transportation networks, and stock replenishment (to name a few use cases).
AI-driven process intelligence platforms can ingest massive amounts of data to discover, monitor and enhance operational and supply chain processes. Some ERP and supply chain software vendors have integrated their core capabilities with acquired process mining companies. Whether or not the data from your RFID system is integrated with ERP, WMS, TMS, etc., process mining leverages AI to transform information into margin improvements.
Data as a Service in High-Volume Settings
At a macro-level, RFID’s highest potential value is realized when systems are deployed early in the supply chain. If you’re reading this as a manufacturer with RFID capabilities, AI’s data appetite might multiply RFID’s value downstream of your current use, especially if you have B2B sales.
For example, you might inform strategic customers of your RFID capabilities, the data it provides, and related AI-driven insights (see prediction #1). This can unlock opportunities such as selling data as a service and improving future contract negotiations.
Take a Page from Walmart’s Book—Move Tagging Upstream
Building on #1 and #2, if you tag purchased inventory—as a manufacturer, distributor, or retailer—try to mimic Walmart’s strategy of having tags applied before they get to you. Even if suppliers have rejected this idea in the past, they may be more willing to deploy RFID systems if you can demonstrate how RFID + AI are improving processes.
Every leadership team should be open to AI proposals, and case studies within their industry are powerful influencers. If your inventory includes brand-name CPG or F&B products, there’s a good chance that Walmart started the conversation and it’s your turn to continue it.
RFID Keep Element in IoT
We should not forget active RFID, especially as part of larger IoT systems. When active tags are equipped with sensors they are an excellent source of information in environments where wi-fi and cellular are impractical. It’s time to reexamine every active tag installation and related data repository to see where AI opportunities are.
Vibration, temperature, and energy consumption are key inputs to predictive maintenance and field service. The same type of AI and machine learning that flags fraudulent credit card activity can alert you to risks before they turn into problems. Add regional weather to get an even better picture of field performance without relying on outdated statistic regression.
OEMs that haven’t already leveraged field data in service offerings need to explore new revenue streams with active RFID in mind.
See Computer Vision for What It Is
Advances in AI are making computer vision systems more accessible. This isn’t a case for buying a VR headset per user; simple photos coming into a trained AI application has enormous potential for quality, maintenance, and even inventory control. That said, it’s not going to replace RFID.
Good computer vision use cases rarely conflict with good RFID use cases, for many of the same reasons that RFID succeeds over barcodes: vision requires a line of sight and RFID doesn’t. Rather than getting territorial about the crossover, such as stock counting, explore using RFID as an AI training partner.
Each AI image analysis has a confidence score—compare what AI thinks something is against what RFID knows it is to improve models. The opposite direction can also work—use computer vision to catch incorrect tag data before they leave your facility and turn into a customer satisfaction issue.
Final Thoughts
The integration of AI and RFID technologies presents an opportunity for enterprises to unlock new levels of efficiency, visibility, and profitability. AI opens doors for businesses to harness the full potential of their RFID-generated data, optimize processes, enhance decision-making, and ultimately drive bottom-line improvements.
As we enter the new year with new budgets and expectations, don’t forget that all of this technology talk has to be part of a wider narrative of optimizing processes to make your workplace and world better for people.