technunes

LAR, doce LAR

Blog Post created by technunes on Jul 9, 2018

If you are wondering what “lar, doce lar” means, this is the exact Portuguese translation for “home, sweet home”. Obviously, the context of this article has nothing to do with the pleasure of being returning home, but the he Lean, Automate and Robotize (LAR) approach can and should be just as sweet and can also avoid some painful experiences in the future.

Before we get deeper into the topic, bear in mind that LAR is the result of my own experiences around automating business processes, and it is not a known term in the community. While I do not have the ambition to push one more acronym into our work routine, I do hope, for the sake of security, efficiency and cost reduction, the approach gets implicitly adopted as a best practice.

There is of course a bunch of articles mentioning that the processes should be optimized before being automated, but with the hype around Robotic Process Automation (RPA), we need to revisit the subject as RPA can also offer a risk for your process automation initiative. After all, your competitive edge hinges not on whether you automate, but how you do it.

So, let’s get some definitions straight:

·        Lean (in a service context) is a process optimization methodology that focuses on improving the effectiveness and efficiency of a process by eliminating activities that do not add value to the customers and the product.

·        Automation of business processes is a technology-enabled approach that is performed to achieve digital transformation or to increase service quality or to improve service delivery or to contain costs.

·        Robotic Process Automation is a computer software or a “robot” to capture and interpret existing applications for processing a transaction, manipulating data, triggering responses and communicating with other digital systems.

An analogy that I like to use when comes to business processes is that a business process is like a road through the mountains and, due to people, systems and procedures, it is full of curves and ups and downs.

The goal of the process optimization is about creating a bridge that will connect the start to the end of the process in the most efficient way. 

In a complex scenario however, this business process can involve ERP systems, legacy systems, office tools and many more. Altogether, this can also offer some change management challenges.

In addition, business will depend on IT to build this bridge and what supposed to be simple, can turn up to be a very sophisticate project, leading to delays, huge implementation costs and frustrations.

As Mike Fitzmaurice would say: in the best of all possible worlds, we would be able to get the requester and the developers a mind meld, but life is not like star trek and mind meld is not real.

The RPA can solve this problem by keeping the process as it is, but let robots to drive the “cars”. Robots can at some level, respond to events just as humans would.

This is all great, but more often than we think, a hybrid approach would be the most appropriate. While we all know that the integrations with an ERP system may bring the complexity of a project to high levels, and many legacy systems does not offer a proper API that can easy out a possible integration, there are parts of the process that can and should be optimized in a proper way. Nintex technologies empowers users to achieve just that, avoiding the IT bottle neck.

The RPA approach should not be considered the answer for everything and a more efficient approach would be the combination of the best of both worlds. For example, by replacing the excel spreadsheet with a proper digital form (mobile enabled) and avoiding the back-and-forth of emails through a proper workflow around a form. 

Lean and automating are very connected and they overlap in many levels because it often involves a process optimization in a digital format. The initiative of automating a process, like introducing a digital workflow, most likely show opportunities to enhance the process itself. In the same way, the initiative of enhancing a process will frequently result on some sort of automation adoption. The robotizing part is connected with the automation because it offers a mechanism of interaction with digital systems in an automatic way, but does not attempt to substantially change the AS-IS process. It is also important to notice that LAR is part of the continuous improvement process defined in Lean.

The fact that RPA does not attempt to change the AS-IS process, gives the idea that the Robotic Process Automation is always an easier, faster and more cost efficient than any other automation approach. This is however not always the case.

What if building the proper bridge did not require IT muscle? That’s where Nintex Technologies come in place.

The scenario below shows the process around receiving an offer, creating a P.O. and receiving an invoice.

·        The requester asks the supplier to send an offer.

·        The supplier sends an offer (email text, pdf, word, excel or even PowerPoint).

·        The offer is validated by the requester.

·        The offer is sent to the internal responsible to create the P.O.

·        The P.O. is created in SAP and sent to the supplier.

·        The supplier will then send the invoice.

This could eventually been addressed by EDI (electronic data interchange), B2B systems or API calls, but it is difficult to expect that all business partners from different sizes, regions and segments can easily adopt such approach. The Robotic Process Automation would then be good candidate to automate this process.

However the “Read document/ extract data” part of the process can be quite challenging because RPA tools does do magic. PDF is a business standard document format and it would be fair to demand the offers to be sent on this format. With that in mind, a robot can be configured to read the offer from the offer repository and using some intelligent OCR and text analysis, figure out who is the supplier, the total amount for the offer and basically transform all unstructured or semi-structured data into structure data. The problem here is that the PDF contained the offer can be formatted in several different ways and it is unlikely that the RPA tool would be able to get the data in a trustful way, unless it attempts to use a 3rd party product like Abbyy to achieve that, which would raise the complexity and the operational cost.

So, let’s revisit the process and apply LAR into it.

At the very start of the process, there is an exchange of email between the requester and the supplier in order to get the offer. The supplier sends an email with the offer attached. The offer document is in a semi-structured format and, as we have seen earlier, it implies some challenges to our automation with the RPA tool.

A way to optimize the process would be to eliminate the email exchange and establish a way to collect the structured data.

Using Nintex Workflow cloud, a business user can easily create an online responsive form to handle the offer submission and the workflow to handle the approval form.

The offer form can be divided into sections. While the Supplier and Offer sections collects the master data for the offer document, the Info and Signature sections support the overall process.

The workflow engine would consume this request and handle the approval of the offer. Additionally it is possible to connect to the cognitive services for text analysis and data classification. 

 

In our example, we capture the sentiment of the comment provided in the Info section. In case the comment is something positive like “20% discount is already applied”, the process will continue in one path of the workflow. If a negative message is provided, like “the offer is only valid until the end of this week”, then the process will follow another path in the workflow.

Using Nintex DocGen, a well formatted document is generated. This document can be the trigger for the RPA robot to start the P.O. process. Alternatively, the Robot can be called directly from the workflow as the orchestrator exposes a web API.

The overall architecture is represented on the image below:

 

 

A solution like the one described in this article is achieved with a no-code approach and can be delivered by business users in a matter of hours. RPA is an excellent technology but should not be considered the answer for all the automation initiatives.

LAR will support the creation of the proper bridges when is feasible and let the RPA do what RPA is good at – to take the robot out of people.

Outcomes