Izem FEKNOUS, Director | Keyrus Management Régions Christelle CHARPENTIER, Senior Consultant, RPA & Finance expert | Keyrus Management In the 1970s, the advent of the microprocessor transformed the way things were produced, and it was seen as the third industrial revolution. Building on the development of computerized production tools, the industrial sector continues to modernize its production lines, to such an extent that we now talk of "smart factories". The shift to the computer age has profoundly affected both large and small organizations alike. It has led their support functions to industrialize their productive chains – especially Accounting – and automate certain tasks, and now they are starting to seek out solutions for automating their processes. Let us take the example of valuing stocks. A stock valuation is arrived at by multiplying the volume of stocks held at storage sites (shops, warehouses, etc.) by the unit price for each type of product. The end amount is obtained through various channels: an inventory of one or more products, the recording of stock entries and withdrawals when the merchandise is received and sold, etc. If we simplify matters in the extreme (one type of piece of data = one database), there are at least four databases feeding into this calculation. The fact that there are many different databases, with different individuals in charge of them (e.g.: the Procurement Department, Finance Department, etc.), very often causes discrepancies, which, in themselves, will affect the nature of the tasks that make up the process. Time that could otherwise be spent conducting business analyses will be taken up by the time-consuming task of sorting out discrepancies. Yet, such analyses are crucial if one is to gain an a posteriori understanding of the business, anticipate changes in it (examples being the seasonality of a product or its geographical nature), and also manage its performance effectively. For such a process to be automated, the databases must be made to "communicate" with each other, unless, that is, you just have one, to centralize the data collected. Such automation also involves redefining, in clear terms, the roles and responsibilities of each individual in the productive chain. To automate is, first of all, to remove sources of errors that give rise to manual tasks, with the aim of significantly improving data quality. In the strict sense, it is about limiting manual action to the initial stimulus that then enables the thing receiving it to perform the full movement. It is not a question of imitating manual tasks, but rather of accomplishing them in a different way, and thus of reviewing and reshaping the processes. We are currently moving from a situation in which we have manual and automated tasks, to one where we are witnessing the emergence of automated or automatic processes, made possible by Robotic Process Automation (RPA) tools. As this shift occurs, the key fundamental factor behind the success of a project to automate a process is the project's capacity to integrate the new RPA tool into the existing setup. Designing an effective process that builds on the existing one What does Robotic Process Automation (RPA) mean? It means the automation of processes using tools, called "robots", that are capable of performing a sequence of tasks. The action by the user no longer consists of performing the task, but is, instead, limited to configuring the robot at the outset and then handling any unusual tasks (ones that cannot be accomplished automatically). According to the principles of automation, the first thing to do is to take stock of the process that is to be automated. Adopting a Lean Management perspective, you need to identify duplicated tasks, those with little added value, the individuals responsible for each task, the "upstream" and "downstream" tasks, etc. However, there are other points to be considered too: • The quality of the data fed into the process at the front end, and coming out at the end of the chain, can affect the performance of tasks. For example, if data are incorrect and/or have come from two different databases and/or are in different links in the chain, they entail what can be a long process to analyze and reconcile the discrepancies in them, so as to ensure they are reliable • What is also required is a survey and assessment of the tools used: am I aware of all the tools involved in my process? Do they enable me to ensure the integrity and quality of the data? Do they help users perform their tasks? To what extent have users appropriated the tools available to them? Do such tools take account of the specificities of the business and/or organization concerned? Once you have fully taken stock of the existing setup, (roles and responsibilities, organization, tools, and processes), you then need to design a target process, along with the roadmap for implementing it, whilst not forgetting to take into account the changing face of the structures with which we work. Thus, if consultants are assigned to work with a high-growth structure facing challenges linked to skyrocketing volumes, the new process will have to "support" such growth in both senses of the word: by both sustaining it, and withstanding it. This necessarily involves designing so-called "scalable" processes, and an RPA tool can ensure such "scalability" of the process. Choosing your rpa tool so as to ensure that the automated process is efficient This choice needs to be based on: • An assessment of how well the tool meets users' needs • The RPA tool's capacity to interface with the entire existing Information System (IS), without limiting the assessment of such interfacing capacity to only those tools now used in the relevant process • The RPA tool's capacity to support the enterprise's growth and strategy • The integration of the tool's costs, including maintenance costs and licensing fees, into the Information Systems Department's budgetary strategy. In as much as the tool can be deployed transversely, it is best to avoid deforming the process to make it fit the tool. If the latter is used in a process that takes automation aspects into account, it then becomes possible to utilize the full range of the RPA tool's technical capabilities. Today, the benefits delivered by an RPA tool are not just limited to automating or sequencing tasks, or to a process. Since an RPA tool can be applied transversely to business functions, it can offer a more overall supervision of processes. It also brings with it a user-friendly dimension that acts to optimize the rate of appropriation of the tool by users. Lastly, from the perspective of steering the enterprise's activity, aside from the time that RPA tools free up – which can instead be used for conducting analyses - , such tools frequently offer a customizable reporting application that can be accessed on any type of device. An automated process: what changes in an organization? An RPA tool implemented either transversely or on a single process (financial, HR, or IT) will provide gains on several fronts that will ultimately make time for undertaking analyses and steering performance: • A reduction in the amount of time that the RPA tool's users spend accomplishing tasks. Their role is henceforth limited to managing one-off situations and establishing the rules of the business functions with regard to configuring the tool • The users, who now take on a steering/supervisory role, can find out the current status of the process in real time, which limits the often time-consuming production of what are sometimes irreconcilable reportings • A collaborative working tool that enables everyone to find out the current status of a process in real time, and to exchange comments on the documents included as attachments • A common repository that improves communication, thanks to the possibility of formalizing and documenting the entire process in the tool, including everyone's roles and responsibilities • A repository that is common to all the users of the process or processes to which the tool is applied, regardless of their individual roles, thus improving data quality by reducing both the adverse effects of having multiple databases, and the reconciliation actions required. ABOUT THE AUTHORS Izem FEKNOUS Izem's 13 years of experience have allowed him to lead projects involving IT consulting, IT-driven innovation, and digitalization, as well as organizational issues and transformations. Izem has developed recognized expertise in the analysis of Information-System-supported business processes and in the area of the governance of projects linked to business functions including Finance, OTC, and P2P. Christelle CHARPENTIER For more than 5 years now, Christelle has been involved in assignments to transform Finance functions, and in particular, in projects to optimize financial processes. She has assisted Industry players in the Energy & Utilities, Web, and Beauty & Retail sectors, and sometimes Banking and Insurance players. These assignments have enabled her to develop her knowledge of financial and accounting processes as well as ERP- and RPA-type tools.