There are numerous articles on the benefits of digitization for manufacturers, but less coverage on common mistakes and missteps to try and avoid. It was a topic that came up during conversations with system integrators at a recent trade show.

The area of expertise for one of the integrators was around robotics, where he noted that there can be a great deal of deployed but underutilized technology. One prime example is welding, where there is a drive to remove humans from “dull, dirty, and dangerous” work. However, in his observation, manufacturers can make critical missteps when implementing new technologies into these applications. The other integrator is more focused on data and analytics across a wide range of end-markets, and noted consistent issues as they work with clients to implement digital solutions.

Automate Before Analysis

Though both integrators deal with different end-markets and technology solutions, both highlighted a consistent first mistake manufacturers often make. Becoming enamored with technology. It can be mesmerizing to watch a welding, bin picking, or warehouse delivery robot. Quickly, the conversation turns into specs, unique requirements, and pricing.

Both integrators highlighted that, if they’re doing their jobs properly, the conversation should go back to business objectives. Maybe the automated robot is the end-goal, but are there more scalable steps that can get you there?

Forgetting the Human

For a long time, machine vision and AI have promised (or threatened, depending on your perspective) to replace humans. That’s often not reality. Over 70 percent of manufacturing processes in the United States rely on a human decision. Even with automation, humans are still making decisions on a manufacturing line, or increasingly based on the data that can now be generated for real-time or corrective analysis.

With that in mind, both integrators highlighted the importance of remembering the human as you look to automate a process. That may be as simple as explaining to an operator why you want to automate a process and providing the necessary training so they can apply their expertise in new ways. In the welding example, the goal is to remove the human from the repetitive work, but still rely on their expert insight to monitor the process and assess results. Without proper communication and training, humans will be quick to dismiss and resist changes related to new technologies and processes.

Overwhelmed with Data

The area of expertise for one integrator was around data and analysis. Data can be like that shiny new robot, and as we generate more statistics, we lose track of the problem we’re actually trying to solve.

In their example, a manufacturer needed three consistent and reliable data points to make a decision. As they automated a process, they learned they could extract more data. More data is better, right? Quickly their analysts were overwhelmed to the point of decision paralysis. It took an honest conversation on the business objective of the automation initiative, and a refocus back to ensuring those three data points were consistent and reliable.

There are areas where more data can have tremendous benefits as manufacturers plan for future automation steps. One of the key challenges with AI is the large amount of data required to train a model. In a quality inspection application, for example, you require numerous images of good and bad products to build a model.

These are usually specific defects, so there’s not readily available synthetic data sources to build and train a model. By digitizing a visual inspection process, for example, a manufacturer can start to gather a database of images and operator notes that can be used for future AI model training.

Plan, Train, and Scale

Considering these mistakes to avoid, one approach for manufacturers can be to digitize first with a view towards automation. Choose a problematic process, and as a first step digitize it. With proper communication and training, your operators gain comfort with new technology.

As a business, you’ll start seeing an immediate return on investment without breaking the bank on a solution that may sit unused on the shop floor. In one example, digitizing assembly directions helped a manufacturer reduce a setup process from one hour to less than 10 minutes. With a scaled approach, you can then gain the data insight to help drive your next automation or digitization decision.