Building Resilience Through Automation, AI, and Data

Aleks Zlatic, ePS

While automation and artificial intelligence (AI) technologies have taken a giant leap over the past decade, their full potential can often feel out of reach for manufacturing business owners. Understanding crossing the chasm between desire and reality when automating label and package printing operations can be daunting. Just ask Aleks Zlatic, vice president of product portfolio and marketing development for eProductivity Software (ePS).

Zlatic has a deep understanding of packaging, printing, software trends and the industries’ evolutions. This positions Zlatic as a thought leader who can offer practical advice for package printing company owners and managers, who want to set their businesses on a trajectory toward success that is resilient despite market challenges, including fluctuating demand and a dearth of experienced, skilled labor.

The Increasing Importance of Data

Emphasizing the foundational role of data in successful label and package printing companies, Zlatic recalls the idiom, “Data is the new gold.”

This statement is perhaps even more true today than it was back in July 2020 when Derek O’Halloran, head of the System Initiative on the Future of Digital Economy and Society and member of the executive committee, World Economic Forum, and Francisco D’Souza, fellow, World Economic Forum, used the definitive proclamation in their World Economic Forum article about using data to generate top-line value. In the article, O’Halloran and D’Souza write, “For companies large and small, around the world, their future depends on using data effectively.”

Today, generative AI dominates the AI narrative, but gen-AI cannot function without accurate and complete data to feed it. Automation for manufacturing only exists when complete real-time machine data is available. Only then can it be used to understand which conditions demand action and how to react. As Zlatic puts it, “Having a data strategy; that’s where it starts!”

Separating the Treasure From the Fools’ Gold

While Zlatic and the World Economic Forum agree on the importance of data, Zlatic also has observed firsthand the struggle surrounding data organization and quality. “A lot of people think they have credible data, and yes, they may have data,” Zlatic says. “But that data is sitting in a lot of different systems, incomplete and fragmented. Further, some data may be useful, while some is not.”

For label and package printing companies, specifically, Zlatic notes, while many converters have good data around estimates and job costing, their manufacturing data is often less reliable. Adding to this challenge is much of the manufacturing production data starts with manual data entry, which is prone to errors, fluctuations, and simply is inaccurate. A human being analyzing such data, let alone an AI learning model, would derive the wrong conclusions.

The first step in turning that around, Zlatic explains, is having healthy discussions at the C-level to develop a comprehensive data strategy. He suggests owners and managers of label and package manufacturing companies answer some key questions: Is it reliable? How do we intend to organize our data?

Investing the time to establish a data strategy is more than purchasing tools and technologies. The process must involve organizational changes with a change management strategy to ensure the new data-centric processes are adhered to across the company.

The Transformative Power of Quick Wins

Once your data strategy is in place and you’ve identified key players in building, maintaining, and refining it, you need to build buy-in from your entire team and develop a change management program. The most effective way to build motivation, Zlatic says, is to deliver quick wins.

Key stakeholders should identify low-hanging fruit with the best data currently available. The second step is to choose an area of the company that can benefit from the new uses of that data, related to a couple key stakeholders. For example, a converter could find that much of its estimating data is exceptionally clean. That converter could use a BI (business intelligence) solution to view this data and start seeing patterns of low margin, low profit customers, etc. Business intelligence systems are ideal for relating data to KPIs and also detecting anomalies and inaccuracies, which are prevalent at the start of any process.

After evaluating the pilot program results, key stakeholders will want to assess the project’s results against the defined objectives and publicize that improvement within the company.

Practical Steps for Scaling Up Automation Projects

Once pilot projects have proven successful, companies can scale up their automation efforts. Although scaling up should be done strategically, there are steps that can help ensure the highest return on a converter’s hard and soft investments in automation.

• Understand Your Starting Point and Goals: Create a comprehensive plan that includes clear milestones, inventory of existing resources both human and technological, and initial timelines.

• Allocate Resources: Determine your budget for personnel and technology to support the scaling process, and then secure necessary resources.

• Expand Gradually: Re-examine timelines and look for opportunities to break up growth into shorter phases. This approach helps manage risks and identify beneficial adjustments based on feedback and performance data.

• Train Employees: Provide formal training to ensure employees are comfortable with new technologies and processes. Then, set up a continuous training and support program with milestones.

• Monitor Performance: Use business intelligence (BI) tools to make the data visual through key metrics, as well as identify areas for data accuracy improvement.

• Gather Feedback: Regularly gather feedback from employees and stakeholders to identify challenges and opportunities.

Once you and your team understand what the data is telling you and can certify it is accurate, applying AI is predictable and fruitful.

Leveraging AI for Automation, Optimization

No article about automation would be complete without a discussion around one of the most powerful tools for automation — AI.

Zlatic understands the apprehension about AI as it seems new and unproven. He wants to assure converters that artificial intelligence technologies aren’t actually novel. “I worked on this when I was a computer science student 25 years ago!” he exclaims.

Zlatic speaks to packaging audience about data, AI, and Industry 4.0 at ePS CONNECT 2024

What is “new” is generative AI (gen-AI). Many of the existing AI technologies fall under the subset of machine learning (ML), Zlatic explains, that ML has been instrumental in the engines behind some of the most productive processes in printing operations, such as automatic scheduling systems. AI’s ability to quickly collect, analyze, and deploy optimized schedules has greatly enhanced converting production throughput and identification of bottlenecks in numerous printing operations.

Furthermore, ML can be used in tandem with gen-AI to simplify users’ interactions using natural language. One example given by Zlatic is using ML to analyze production data and identify bottlenecks in a specific area of label and package manufacturing using a series of natural language prompts. Once these bottlenecks are identified, gen-AI can even be used to create new scripts to drive standard operating procedures.

“Heuristic AI adds a powerful layer of optimization that today’s technology and automation bring to manufacturing,” says Aleks Zlatic. “Modern software solutions leverage AI-driven algorithms to dynamically balance constraints such as deadlines and resource availability while optimizing efficiency. These technologies turn raw production data into actionable insights by instantly adapting schedules, modeling 'what-if' scenarios, reducing make-ready times, and enabling manual overrides only when needed. This ensures manufacturers can operate with maximum productivity and minimal downtime, even in the most complex environments.”

Harnessing the Full Potential of Automation, AI, and Data

The most efficacious optimization resource a converter can leverage isn’t a form of technology — it’s identifying strategic partners committed to uncovering the full potential of automation, AI, and data collection and analysis. These partners should understand the strengths and limitations of all subsets of automation and AI technologies and the needs and opportunities for printing businesses to leverage these technologies. These partners should also have experience developing data-driven technologies.

As the packaging operating unit of eProductivity Software, ePS Packaging can tap into its 30-plus-year history of being the leading global software provider for all segments in the printing industry. The company’s powerful business and production workflow automation technologies have been proven to reduce the cost of sales and manufacturing and to grow revenues for packaging and printing companies around the globe.

Part of ePS Packaging’s success is its consultative approach, which ensures converters receive the support needed to navigate the complexities of manufacturing optimization. This approach is paired with a comprehensive suite of software products addressing the needs of printing companies, from estimating and production planning to real-time data collection and workflow automation.

To learn more about how to work with ePS Packaging, visit epackagingsw.com.