The cost, by which we mean the price the client pays for automation of processes, is a result of the assumed billing model, e.g. time & material or subscription-based service, in which the rate is correlated with the number of saved FTEs or the number of work hours equivalent performed by robots in a unit of time.
Cost calculations derive from a number of “source” costs incurred by every implementation company and can be summarised as follows:
The table presents the existence of three cost blocks: project setup, cost of development of the solution together with licences, maintenance and support, and the cost of change request implementation.
The middle block is the most cost-generating one, as it includes activities performed over many days or even weeks by highly specialised technical personnel and licence block.
Thus, there are many factors that affect the cost of robotization, often calculated as TCO – Total Cost of Ownership, however, without juxtaposing it against the expected benefits, meaning the scale and effects of automation, such robotization cost by itself carries little informational value.
There is a lot of truth in saying that when implementing new technologies, such as in the class of robotic process automation, you need to be ready to spend tens of thousands to earn hundreds of thousands, and spending hundreds of thousands can earn you millions.
It is a frequent model of cooperation between a technology user and a technology supplier and it includes selection of two or more RPA-class technologies. In our case, we use two low-code tools – UiPath and Automation Anywhere – and G1ant that requires programming skills. Vendor-agnostic means having more than one robotics software supplier and is a derivative of a business model of strategy of managing automation projects. Having more than one supplier is tightly correlated with maturity of the robotics center and directly stems from the established rules of managing such a center, in which high utilization of bought licenses is extremely crucial as the number of automated processes and the number of tasks related to their maintenance grow. Another argument for the vendor-agnostic model is related to the need for selection of technology suitable for the IT competences in an organization. If automations are created by developers from IT department, which happens in half of Polish companies, then there is a natural tendency to select tools created in languages with similar syntax as those commonly used in programming languages. If an organization goes for dissemination of automation and so-called citizen development, only low-code tools like UiPath and Automation Anywhere give a chance of successful implementation. A deciding aspect in technology selection is also security. RPA tools differ between themselves in terms of components and the way information flows between them.
Simple automations have become a standard and working with structured digital data is not a challenge anymore. Suppliers of RPA-class technology intensively develop their products in the area of Cognitive/Intelligent Automation, which means that they enrich RPA with elements of artificial intelligence (AI). It mainly refers to machine learning with initially trained models (algorithm-based) for working on unstructured data, data included in robust files such as lease agreements, purchase invoices, requests for settlement payments. This technology will not only obtain data essential to decision-making, but also analyse it in order to eliminate possible misuse. Companies that use such a platform feed models with their own data, so that they teach them further to allow reaching impressive levels of so-called success matrices. The future will bring linking robots with both Natural Language Processing (NLP) technology, so that voice as a stream of input data as recognised, understood, as well as with sensors which will allow machine and device usage prediction, and failure prevention by advance replacement of machine parts. It is predicted that as digitization of business processes, markets and whole societies advances, the share of usage of RPA-class tools will grow.However, automation itself will be dominated by AI-based technologies, whose share in process automation will spring from 5% today (2021) to dominant in 2025. This is what Phil Ferst of HfS had in mind in 2019 when sharing his thesis that RPA is dead.
Numerous research on innovation absorption in organizations proves that available technologies change very fast, while organizations themselves do definitely slower. This imbalance in development is a significant matter as it has always accounted for – and will still account for – the competitive advantage of market players, economies and societies. The essence of this imbalance in the pace of change is inertia and unwillingness to change caused mainly by the fear of the new. Resistance toward change is much more common than willingness to make radical changes and it is a phenomenon that accompanies people in all latitudes and time periods. The driver of any change is always an individual that exhibits traits of a leader, who can inspire others and thus initiate the process of change. Process automation is not solely a technological matter as it requires change in organizational culture, planning and carrying out educational activities, helping in acquiring new practical skills and helping those employees that are under threat of technological unemployment. Ed Gabrys, who is an analyst in a research company Gartner, stated that we move from the world in which people work like computers to one where computers work like people. Automation has an impact on our lives – starting with our workplace – which will only continue to increase. The observed acceleration of digitalization and automation caused by the Covid-19 pandemic is not the goal in itself. Consciously or not, we approach the development phase called Industry 4.0. Thus, it is really worth paying sufficiently much attention to process automation, defining its goals, and above all building competencies not in the area of technologies themselves but also in leaders who notice what is needed and take action.
The choice whether to go for building own competences or inviting a specialised agent is a matter that each company needs to individually assess and then make a decision. Of course, going the path of building own competencies does not prevent one from sourcing services from outside of the company, and building competencies can begin with starting cooperation with an external entity. There is also a third model: commissioning automation of 20 first processes by an external company, so that after a certain period of time (usually ca. 2 years) and mutual agreement, the built team can be absorbed into the structures of the commissioning company. Incubator service obviously includes all necessary regulations.
Building own competencies in the area of RPA or Intelligent RPA is like building a new branch of a company from scratch. It is a daunting and costly task, yet it can be profitable with the right size of the company, rigorous towards the matters related to RPA Governance, and large internal potential for automation. One of the first actions is selection and assessment of robotics platforms, which will fulfill requirements now and in the future. A solid technology vendor assessment together with negotiating satisfactory licencing conditions are done simultaneously with filling posts crucial for proper conducting analysis, development, testing, implementing and maintaining robots. In the beginning, predominant is a model where roles are combined due to high cost of technological personnel and a relatively low volume of tasks. It should then be ensured that a well-considered concept for change management is created as technologies evolve fast, while organizations slowly, which causes many problems.
Basing on known bet practices, a robotization process requires 7 or more distinct roles:
Head of RPA (Process Automation – an experienced manager who understands the business and technologies, manages the whole center, the team’s effectiveness, projects, and cooperation with the overseeing body, e.g. RPA Management, who realises the strategy agreed with the company’s headquarters.