Zebra Technologies AI Leader Advises Companies to Rethink Their Frontline Data for AI
Zebra Technologies Corporation, a leader in digitising and automating frontline workflows, is advising businesses to make better use of their frontline data to maximise the benefits of AI solutions.
The helpful advice from Stuart Hubbard, Senior Director of AI and Advanced Development, Zebra Technologies, offers hope following a series of reports, media headlines and statements from tech leaders saying we are running out of high-quality data needed to train, test, and advance AI models.
According to one report, researchers have observed an increase in data use restrictions among web domains used for generative AI training datasets. Between April 2023 and April 2024, 25% of data from the highest quality sources has been restricted, as websites limit data scraping for AI training. Some argue that the stock of human-generated public text will be used up between 2026 and 2032, or earlier.
“The quality of data poses another challenge for generative AI and solutions using deep learning,” said Hubbard. “Inaccuracies, biases, or a lack of variety reflecting real-life use cases in the data will lead to poorly performing AI models that do not support business growth, intelligent automation, or a more connected frontline.”
To address these issues, new AI models, data capture, and analysis solutions are being developed which open new avenues for data sources that are currently underused. One such untapped resource is business data generated on the frontline—data produced in real time within workflows, on devices, and through industrial technology systems.
Large volumes of data generated at the intersection of workers and machines offers a rich yet largely unexplored landscape for industries such as retail, manufacturing, and logistics. Frontline data can include:
In manufacturing: defects and anomalies in materials, components and finished items, packaging quality, labelling, barcodes and characters, quality of returned items.
In logistics: parcel volumes and shapes, vehicles available, delivery routes, delivery preferences by customer profile and geography, proof of delivery photos, temperature and humidity.
Technology companies like Zebra are spearheading efforts to digitise and intelligently automate frontline operations, enhancing both the availability and quality of enterprise data generated in industry environments.
This approach increases the volume of relevant, reliable, timely, and accessible data. And developing advanced data integration and processing tools alongside robust ethical guidelines and legal frameworks can ensure responsible AI development.
“Business leaders in IT, operational technology, and AI development should focus on capturing and harnessing their frontline data by employing technologies like mobile computers with computer vision, agentic AI, and voice AI, and smart cameras, sensors, and software with deep learning and 3D algorithms. By doing so, they can transform this raw data into actionable insights and value,” said Hubbard.
“Additionally, well-established technologies like radio frequency identification (RFID) can be viewed in new and promising ways. Mobile and fixed RFID solutions can swiftly capture large volumes of highly accurate data which can be fed into AI models and systems for training and analysis for quality, planning, and process optimisation, and support digital twin development and scenario planning.”
As workflows become increasingly digitised, mobile and edge devices generate their own operational data that offer insights into process efficiencies, productivity improvements, and resource utilisation.
This operational and contextual data, including user interaction patterns and environmental context, provide a deeper understanding of environments and behaviours. It can also enable predictive maintenance using machine learning, by identifying equipment anomalies before they cause disruptions.
For more information: www.zebra.com
