In this second article of our series on AI’s impact on business decisions, we will explore the value and limitations of process automation technology. For decades, businesses have used rules-based process automation to automate manual tasks and assist with decision-making. With our business environments becoming increasingly complex and data overwhelming, AI appears to be the solution to every problem. But is traditional process automation technology obsolete? Read on to find out what its use cases are and what its limitations are in the age of AI.
The Rise of Traditional Automation Technology To Support Business Decision-Making.
Decades ago, to alleviate the challenges associated with human decision making, businesses turned to computer and process automation technology. In particular, this technology allowed for repetitive tasks to be carried out more quickly and accurately than would be possible by humans alone. As a result, operational efficiency has improved significantly. Also, by businesses automating repetitive, simple tasks has enabled employees to focus on more strategic tasks that demand creativity and complex problem-solving skills. As a result of businesses implementing computer and process automation, most business operations consider process automation technology an essential component of successful business operations in the digital era.
Even in the age of AI, process automation is still key for increasing operational efficiencies, So let’s dive into exactly what process automation means. Specifically, process automation is procedural oriented and works off simple business rules to process data transactions or to trigger alerts. Examples of this type of software include automated file transfers, email automation, automated event log monitoring, automated batch processes to name a few. For more examples, See Forta’s 15 Examples Of Business Process Automation for more details.
How Does Process Automation Technology Help Decision-Making Processes?
Though some of the process automation software appears simple in the age of AI, it works. Even with advances in AI and data science, I would expect that process automation software will still continue to be a useful tool to support business decisions for many years to come.
Also, to best understand how best to use automation for decision-making, it is best to use some conceptual frameworks. Gardner has documented some great frameworks for both decision automation and domains of complexity (Cynefin framework). Specifically, these concepts help to identify what type of automation to best use as well as identify the degree of complexity of the problem. To detail, see the following two sections for how process automation technology can provide different levels of decision support within varying levels of complexity.
1. Three Levels Of Automation Technology To Support Business Decision-Making.
According to Gartner, there are three levels of automation when it comes to supporting decision-making. To detail, See below for a description of each of these levels of automation and how process automation can assist problem solving.
- Decision Automation. This type of automation is usually used to automate simple business processes. So process automation is well suited for providing decision automation on simple tasks. However, businesses now can increasingly leverage AI and data science for more complex autonomous tasks. Usually in these cases, the automation makes decisions using prescriptive analytics or predictive analytics. Also, Its benefits include speed, scalability and consistency of decision making.
- Decision Augmentation. With decision augmentation, the system recommends a decision, or multiple decision alternatives, to humans using prescriptive or predictive analytics. Further, decision augmentation benefits lie in the synergy with human knowledge and data repositories. To explain, business problems are becoming more complex and data intensive. As a result, AI and data analytics software is better suited to rapidly analyze high volumes of data and to deal with complexity. So except for very simple tasks, process automation can only provide rudimentary decision augmentation.
- Decision Support. Here human employees make the decision, supported by descriptive, diagnostic or predictive analytics. Thus, the main benefit of decision support lies in the combined application of data-driven insights and human knowledge, expertise and common sense, including “gut feel” and emotions. In this case, process automation may be valuable for periodic batch output to supplement decision support. However, AI and data analytics tools can more likely do better as well as providing real-time support to making decisions.
2. How Well Can Automation Technology Support Decision-Making In Complex Environments – The Cynefin Framework.
To better pick the right automation to support decision-making, it is important to identify the complexity of the task or environment that the automation will work in. Thus, this is where the Cynefin Framework can help. Specifically, this framework defines levels of complexity from simple to chaotic. Surprisingly, all levels of automation (process automation, data analytics, AI) can work in these various levels of complexities. However, different types of automation tend to work better or worse for varying levels of complexity. Specifically, see descriptions below for each level of complexity in the context of automation and human decision-making.
These situations are stable and predictable, and operate according to clear cause and effect. Examples include payroll processing or call center routing. In many cases, process automation works best and most effectively in these situations.
These situations require expertise or analysis to identify cause and effect, often using expertise with a known problem-solving practice. Examples include insurance fraud, asset management and marketing campaigning. Further, this level of complexity may be better suited for data science and AI.
These situations involve multiple relationships and interdependencies. Further, the decision-making process requires an effective analysis using a systemic or holistic approach. Specifically, the decision-making process can use simulations to see how decisions can affect far-flung elements. For example, analysts would use simulations to model “what if” scenarios such as supply chain disruptions. So, in complex situations the human is definitely in the lead, but decision makers can use AI and data analytics to augment or support the process. In comparison, businesses can and do use process automation, but it is limited. In most cases, businesses will use process automation as part of periodic batch processing.
These situations have unknown causes and effects, with unclear or dynamic interdependencies. As a result, small changes may have seemingly disproportionate impacts. Also, decision making is very difficult and requires experimentation and learning by doing. Examples include stock market crashes, battlefields and natural disasters. In this situation, humans definitely are in the lead and may be supported with a full range of decision support automation.
For more information on degrees of automation and level of complexity in decision-making, see Gartner’s Would You Let Artificial Intelligence Make Your Pay Decisions?
6 Limitations of Traditional Process Automation Technology in Decision-Making Support.
So to summarize, process automation is still a viable tool in supporting decision-making even in the age of AI. However, it does have its limitations. See below for a list of 6 key limitations with process automation.
1. Limitation Of Rule-Based Systems.
Rule-based systems rely on a set of predefined rules to make decisions. However, these systems have a limitation in that they cannot learn from new data or adapt to changing environments.
2. Needs Human Intervention To Adapt to Changing Environments.
Traditional process automation systems are designed to work in a specific environment and with specific data inputs. So they are not able to adapt to changes in the environment or new data inputs. Thus, these systems require human intervention to handle errors and exceptions.
3. Cannot Support Complex Decision-Making Situations.
Traditional process automation systems are limited in their ability to support complex decision-making situations. In particular, they are designed to follow predefined rules and cannot handle situations where there is no clear rule or where multiple rules conflict with each other.
4. Not Able To Deal With Ambiguity and Uncertainty.
Traditional process automation systems are not able to deal with ambiguity and uncertainty. Undeniably, they require clear input data and rules in order to make decisions. Further, when faced with ambiguous or uncertain data, these systems fall short.
5. Limited Capability To Collaborate With Humans.
Traditional process automation systems have limited capability to collaborate with humans. Namely, process automation technology is best used in a batch process and has only limited ability to interact with humans in a meaningful way. Further, this technology is limited in its ability to take into account human input or feedback when making decisions.
6. Not Suited For Large Data Sets.
Also, traditional process automation systems are not suited to work well with “Big Data”. This is because they are designed to work with specific data inputs. So due to the increasing amount of data that is generated today, a whole new scientific field, data science, is now established. Namely, businesses are now more and more leveraging data science over traditional automation that cannot handle massive data sets in real-time.
For more on our Series on AI Impact On Business Decisions, see
- The Human Problem-Solving Process – Part 1
- Process Automation (This article) – Part 2
- Data-Driven Decision-Making – Part 3
- AI Impact on Business Decisions – Opportunities (To Be Published) – Part 4
- AI Impact on Business Decisions – Limitations (T- Part 5
Greetings! As an independent supply chain tech expert with 30+ years of hands-on experience, I take great pleasure in providing actionable insights to logistics leaders. My background includes implementing 100s of innovative solutions using emerging technologies and a data-centric development approach. I have also provided business intelligence (BI) solutions for 1,000s of shippers. For more about me, click here.