The well-known fear of change has already been mentioned in another post. Digital dead-end thinking definitely also belongs to the category: emotional attachment to the technological current state. However, the explanation of opposing new technology and automation processes is much more precise. In the case of document digitization in general, and specifically for the digital capture of safety data sheets (SDSs), it sounds something like this:
On sds automation
“Automation only makes sense for us if it is 100% accurate!”
Let’s think the statement through. It means someone doesn’t accept a fax in the 80s because e-mail doesn’t exist yet. It also means that someone decides against smartphones because they still have to operate the touchscreen manually. This person wouldn’t have gotten off the horse for a long time, because cars don’t drive fully automated, interventions are still necessary, at least until today.
You see, it becomes quite absurd. Many technological advances were not possible without intermediate steps.
Underlying all of this is something else entirely. Companies want to bypass the intermediate steps. They want to get to the shortcut that can’t exist. As long as it doesn’t exist, they prefer to wait. It’s a costly decision. Artificial intelligence (AI) or machine learning (ML), in particular, are now hard to imagine any industry without. Those who wait until 5 years from now to tackle these important AI decisions will probably be 10 years behind. This is because the new processes must also be rolled out in the company. The digital corporate culture is part of the process. Employees, after all, also have to be brought on board. All of this takes time. The sooner you start AI digitization, the more beneficial it will be.
Anyone who assumes digitization will be easy, who believes there is a perfect or magical path to automation without detours, is sorely mistaken. Perfectionism is the ultimate brake on innovation. Digitization is like business: you have to start somewhere at some point.
I think there is also a massive information gap between what automation or AI is capable of doing and what it is not. Hence the false expectations.
AI does not decide between true and untrue. AI can weight the truth of a statement with values between zero and one. For this purpose e.g., historical data is used to train the model. For example, a statement by the AI system of 0.95 would mean that the statement is true with a probability of 95%.
AI without a human is therefore often not useful. The human must supervise the system at all times. At SdbHub, it’s no different. We develop these digital assistants to help professionals. At no point does the human lose his or her decision-making authority. The human just handles it much more skillfully and purposefully.
From past implementations, we see the following promising patterns in terms of digitalization:
- The automation solution should be seen as part of an overall digitization strategy.
- The introduction of digital assistants for automation, such as SdbHub, requires an iterative approach.
- Digitization and the introduction of AI solutions is a process that happens gradually and not overnight.
- Solutions like SdbHub not only enable companies to look at existing technical possibilities when interacting with documents. They broaden horizons and create awareness for new use cases.
- After implementation, the automation solution should be continuously improved. In this regard, user feedback is essential and should be considered continously.
