Robotic Process Automation (RPA) is a growing technology around the world that makes it very easy to develop, deploy and manage the automations across a company. These robots emulate human actions interacting with a variety of systems and applications. They can perform actions like reading and identifying what is on a screen, clicking and filling out a form, sending an email containing a report to stakeholders, comparing client signatures, extracting information from a scanned document, and much more! RPA and AI are very friendly, so many AI techniques can be incorporated into the automations. This allows us to associate them with Machine Learning models and OCR skills creating a robust and powerful tool.
You are probably wondering if your process can be automated because you are aware of some scenarios where you think the automation is probably not feasible. That is not a problem at all. The robots are able to identify the different scenarios and execute the actions for what is needed. For those scenarios where the automation is not feasible or is not worth the development, reports can be created alerting the business team about them.
– Data: when creating the automation, multiple reports can be created to identify and classify the data that are being processed by the robots, as well as the different scenarios and results that can be stored in a report format. That allows us to keep everything on track and use this information to analyze the data. All this information can be used to assist managers in decision-making. Normally, most of this information is not stored and tracked when performing manual analysis by a human.
– Data Security: Data leaks are very common nowadays, and most companies are concerned about it and investing millions of dollars to improve their security. The risk and the repercussion of it can be devastating for many businesses. Since well-developed, following good practices, using RPA can reduce the risk of leaking sensitive data by reducing the number of people interacting with this data.
– Quality: Standardizing a process and automating it brings up some big advantages. Robots neither mistype a message nor forget to do the math before applying the correct amount into a given system. This ensures that the process will be done according to what was designed and can save your company from the cost of poor quality.
– Efficiency: It is probably one of the biggest advantages when we talk about robots. Robots can collect data, process, and predict events with a faster speed and accuracy. A robot doing repetitive tasks allows employees to do anything more creative and productive. There are many processes that can be done 1000x faster, and without breaks, 24 hours per 7 days.
– Capacity: When an RPA project is designed to run in parallel, running multiple robots can improve the company’s capability to handle demand fluctuation. It is very hard to allocate people to a process that they are not used to working with. On the other hand, a robot is just a virtual machine running an RPA project that reflects a different process. Therefore, if more capacity is needed for a given project, it is possible to create and run more virtual machines or even allocate them from an idle process.
Mapping a business process is an important job when it comes to process automation. A few details need to be considered when someone will map a process:
A Business Process Mapping (BPM) usually takes the form of a flowchart when separated on a case-by-case basis in how different work items should be processed. A flowchart makes it easier to develop an automation when you know exactly how to handle exceptions and specific situations.
Other benefits of mapping and keeping documentation of your process are: improve Communication, Visibility, Standardization, Auditing, Training, and Continuous improvement. Also, with the map of your process, you can:
Many processes that are not mapped remain only in people’s manual routines and thinking. In these cases, it is very important to review the process mapping a few times, instigating these people in an attempt to map 100% of it. Inefficient processes cost people time. Bottlenecks, complexities, and lack of understanding can hide improvements. Mapping allows for better visibility into business operations, helping to create efficiencies at scale.
It is also relevant to assess the maturity of the process being mapped, as well as the mapping process to obtain an estimate of the level of optimization results. The lower these maturities, the more often repetitions and mapping revisions, reworks in automation developments, should be expected.
There are many ways to start the development of an automation after the BPM is built. The sequence below is a general template of development to follow when it comes to testing the automation and aligning expectations with the business:
– Review the BPM
– Align the goals of the automation
– Review the last improvements and changes in the process so a bot can do the job.
– Build the automation using an RPA software or technology
– User Acceptance Tests (UAT)
– “Soft” Launch (proof of concept with limited runs
– Stabilization and Maintenance transition
– Maintenance of the business process for how long it remains intact.
Now that you know what RPA is and how important it is the role that BPM plays in it, it is time to find out the best (and worst) scenarios when it comes to implementing a Robotic Process Automation solution.
The first thing to consider when thinking about developing a bot to automate a process is how much effort/time it usually takes to do this process manually. If a process usually takes too much effort and/or too much time to be done then it is probably worth implementing a solution to automate it, even if the effort to develop the automation is high, considering compensation in the medium- to long-term. On the other hand, it is not advantageous implementing an automation for the opposite case, where the effort and time spent to build an RPA solution would be much greater than the effort/time to actually do the process by yourself. Of course, for both scenarios above, it is crucial that you know how often the task needs to be done and take that into account when it comes to deciding whether or not to automate a certain process.
Another factor to keep in mind is how your process works. Is it rule-based or is it guided by subjective decisions? If it is rule-based, then an automation process is probably a good fit because it can easily be translated into programming logic. However, if your process lies on subjective decisions, it might not be a good idea to implement RPA. Of course, that this may be the case for attended automation, nevertheless, the very main core of the process should be rule-based.
It is also important to take into consideration what type of data your bot is going to handle. Software bots excel when it comes to extracting and manipulating structured data (tabular data). However, dealing with unstructured data (text, images, videos, audio) may be much harder. This is not a deal breaker, though, it is possible to implement RPA solutions that can successfully process different types of unstructured data, especially with the integrated use of Machine Learning. The important thing is to reflect how viable such implementation would be considering the previous criteria.
Since RPA is a layer that sits above different systems, processes that use different systems at the same time would be a good fit for implementing an automation. It is possible to automate processes that use only one application, however, there would not be 100% use of the benefits of RPA, and in this case, the implementation of a macro would be enough.
Last but not least, it is crucial to know how often the process may change. If your process is constituted of many tasks that are constantly changing, then you should definitively not try to automate it. Imagine having to frequently adapt and change your automation to fit a newly and just modified process, does not sound very good, right? In this case, the effort and time spent in developing and frequently modifying your automation is going to be greater than the effort and time saved by the use of the automation itself.
To identify where RPA can deliver highly-relevant results, processes should not have blind spots. In other words, they must be very well-defined, mapped, and standardized. When you start automating a process that is not entirely clear, you may find out during the development that some steps cannot be automated in your company and that may mess up the original automation architecture, project estimation, and savings.
With that in mind, companies might have multiple processes that can be automated with a variety of manual tasks that are performed daily. Some of them are composed of multiple repetitive activities of copying data from multiple sources, verifying its integrity, applying rules to it, sending e-mails with this data, or pasting it on a different system, and these are exactly what RPA is aimed at. However, all steps performed in these activities must be well-defined and static over time. Moreover, in large companies and corporations, some process activities that must be error-free might be overly performed by multiple people in order to achieve a lower level of human error. These activities are also perfectly suited for RPA, since automation may drastically reduce ETA and do it in a more reliable way.
Processes with repetitive actions and high transaction volume tend to benefit a lot from RPA. However, these processes might be very complex, and they may cause business issues when trying to automate them as a whole. Instead, you may want to consider breaking huge complex processes into sub-processes and start automating them in steps. This way you start creating value as early as possible and adapting the culture of the company to the use of RPA. Besides that, processes that need human intervention for validation or any other purposes could also use this strategy so that a human can validate the transactions between sub-processes with a minor impact on processing time.
Furthermore, some processes may have multiple transaction scenarios to be processed and each one may have to be processed in a different way. That may add a lot of complexity to the automation. Adding business exceptions for special transaction scenarios that are not very frequent may be a workaround to reduce development time and process complexity. This way, special transaction scenarios would have to be manually processed.
As observed during this article, there are a lot of pros when automating a process using RPA techniques. There is a wide range of possibilities, as a lot of tasks can be automated from a simple spreadsheet data extraction to an OCR PDF extraction. However, there are a lot of things to consider before starting the implementation and it is very important to be clear about RPA boundaries and process mapping. When well designed, RPA is a powerful tool for your company with amazing outcomes.