How AI and machine learning can revolutionise labelling software

Gurdip Singh, CEO of Kallik, discusses how artificial intelligence (AI) and machine learning (ML) can be incorporated into labelling and artwork management (LAM) software to prevent errors, enhance the user experience and give companies a competitive advantage. 

With an estimated 50% of pharma and medical product recalls caused by labelling errors, it’s evident that some businesses need to review their label creation methods to ensure accuracy and avoid falling foul of the regulator.

The importance of accuracy 

It goes without saying that the accuracy of product labelling is crucial to avoid product recalls and the hefty fines that can come with them. On average, these fines can cost companies millions of dollars, depending on the scale of the recall.

In an industry built on trust and safety, this too could hamper a brand’s all-important reputation – something that can take years to rebuild.

Risk and regulation

Should a company identify a label that needs amending, either due to an error or a change of regulation, manual methods are slow and unintuitive. In fact, it can take 10 humans weeks to change 1,000 labels. For those selling millions of units of a product each month, in various territories, this timescale simply isn’t good enough.

Combine this with the fact that labels typically need changing up to five times a year due to updates, new requirements, or regulation, it’s easy to see how manual label creation can cost companies significant amounts of revenue.

There’s also the added risk of human error, which is an unfortunate yet sometimes unavoidable aspect of labelling. Given these financial and reputational risks, it’s perhaps surprising that many pharmaceutical giants continue to use manual methods of label and artwork creation.

To highlight the potential severity of mislabelling, in 2021, DeRoyal Industries recalled 138 of its surgical procedure packs, due to an error where 1% lidocaine was mislabelled for 0.5% of bupivacaine. The FDA identified this as a class one recall, essentially the worst-case scenario, meaning that there could have been a potential threat to life if these packs were used for their intended, and labelled, use.

The role of new technologies 

With so many moving parts, it’s important that manufacturers utilise the technologies available to take away some of the pressure that comes with labelling.

Figures from Gartner have revealed that, by 2025, 85% of businesses will have adopted cloud-based technologies into their day-to-day operations. With huge benefits such as connectivity, accessibility and scalability, there’s little doubt that the cloud, artificial intelligence (AI) and Machine Learning (ML), will play a crucial role in revolutionising the way in which the labelling industry currently operates.

By using digital technologies to create and manage the LAM process, the aforementioned risks can be mitigated, with the labelling process streamlined. Combine this with the introduction of AI and ML, businesses will notice significant performance uplifts.

Digitising Labelling and Artwork Management (LAM)

While the use of AI and ML is still relatively new within the LAM market, experts are crediting this slow adoption to the concerns a potential mistake could cause to human life. However, without the traceability and visibility that automated software provides, these mistakes could be a lot harder to identify and much slower to resolve.

With that in mind, there are engineers and programmers in the labelling sector who are working behind-the-scenes to find innovative ways to utilise AI and ML to remove those risks and fine-tune their capabilities.

In 2022, Kallik partnered with Aston University to enhance its offerings by integrating AI and ML into delivery and migration of Kallik’s LAM software. 

Academics at Aston University have previously worked AI and ML algorithms into solutions that can be embedded into pre-existing software, helping businesses stay at the forefront of AI integration – yet, this is the first instance of AI being used in the labelling sector.

Kallik’s LAM software is a cutting-edge platform employed for safety-critical purposes, guaranteeing uniformity and assisting adherence to the most recent legal, regulatory, marketing and manufacturing standards in all aspects of packaging and labelling.

This software breaks down the label creation process into five key stages, supporting all aspects, from organising pre-approved assets, or the building blocks of a label, to the product’s distribution.

Step-by-step advantages

Through this Knowledge Transfer Partnership (KTP), Aston University was able to share with Kallik expert academic insight into how AI and ML works, specifically helping to enhance the artwork creation stage of the labelling process by simplifying it into an automated procedure.

During the artwork creation stage, the elements of the artwork combine, offering users a visual example of the finalised label. With the insights offered from Aston University, Kallik’s LAM software now presents two options for expediting artwork generation: Automated Artwork Generation and Cascade.

Automated Artwork Generation (AAG) dynamically constructs the selected pre-approved content. Through the utilisation of intelligent, pre-approved templates, the artwork is generated autonomously, removing the need for human involvement. Whereas the Cascade element allows artwork designers to exercise complete control over the usage of label components by seamlessly channelling the approved content directly into Adobe InDesign or Illustrator, ensuring designers have full visibility of the whole process.

Together, AI and ML have been strategically used to provide clients and customers with unparalleled capacity, empowering them with the ability to choose between a semi-automated or fully-automated artwork creation process. 

We recently worked with a company in the regulated market, who benefitted from the LAM’s automation, helping to quicken its product’s time to market by as much as 50%, while also reducing the potential for human error.

Ascertaining the competitive advantage

Compounded with a higher time-to-market, wide-scale product recalls and dramatic blows to reputation, it has become increasingly clear that using a manual labelling system carries far more risk than an automated one. This is especially true when considering that operators in the medical device and healthcare sector have built their reputations on trust in their brand, so any mistakes could severely hamper this consumer relationship – not to mention any potential risks to human life.

An increasing number of businesses are discovering the advantages of adopting a unified, all-encompassing LAM solution that comprehensively organises and streamlines the entire labelling process. Already, AI and ML has been used to ease the risk of human error and shorten the product time-to-market, taking the average waiting time for labels from weeks to seconds, significantly helping manufacturers ascertain a competitive advantage and stay at the forefront of technological innovation.

Source link

This post originally appeared on TechToday.