Reading Time: 22 minutes

Ambitious RPA initiatives often end up as isolated pilots in scattered business silos, and the dream of cross-organizational process automation remains a distant memory. This is not necessarily due to the technology used; rather, it’s due to a lack of governance. How do you manage individual RPA projects and how do you manage RPA enterprise-wide? 

Why RPA governance? 

Robotic process automation (RPA) is a fast, non-invasive automation method that builds on the existing infrastructure and system landscape without changing it. It is an intuitive, low-code technology that even citizen developers in business departments can utilize to automate their own business processes. 

latest report
Learn why we are the Leaders in API management and iPaaS

So is RPA easy? That’s just one side of the coin: the great promise that initially led to a downright hype and run on RPA, only to disappoint the high expectations in terms of enterprise-wide automation in the medium term. A study in 2016 found 30–50% of all RPA initiatives failed. Two years later, another survey showed that while 53% of participating companies had long since embarked on their automation journey, only 3% had been able to scale their digital workforce enterprise-wide

The end of the automation journey: why many RPA pilots stay stranded on desert islands 

RPA is easy, there’s no mistake about that – but it’s only half the truth. Because the other side of the coin is this: RPA, intended as an enterprise-wide project, is complex. Automating tasks here and there? That’s done quickly. But orchestrating process automation across the org, scaling it, and digitally transforming companies? That’s a much more demanding challenge. 

Initial RPA pilots start in several business departments and achieve quick, modest successes in these areas. However, the automation journey regularly stagnates at this stage as well: 

RPA […] automates specific tasks in other words, mimicking human behavior at the level of individual work. This has a significant impact at that micro level, but leads to isolated improvement for the macro organization. [These] “islands of innovation” [are] not bad per se, and can actually make sense as a start on a digital transformation journey, but until you connect those islands, you aren’t going anywhere.

Organizations should consider stepping back and questioning what they meant to achieve in process automation (and if those goals were met). If it was the automation of a specific task, then, by all means, proceed with RPA. If, however, it was a broader goal of digital transformation, then RPA used in isolation will not get you there.

Gustavó Gomez

How do you turn scattered “RPA islands” into a wide, cohesive automation landscape? The answer is governance. 

Steering RPA the right way: governance will guide you to success

Governance [Latin: gubernare: to lead, steer, direct] originally stems from the field of state and government, and when applied to the business sector, means in the broadest sense the overall supervision of organizations or projects. RPA governance is situated between IT governance and project governance: 

  • IT governance (ITG) is defined as the processes that ensure the effective and efficient use of IT in enabling an organization to achieve its goals.
  • Project governance is the set of policies, regulations, functions, processes, and procedures and responsibilities that define the establishment, management and control of projects.

RPA governance supports effective IT business alignment and the selection and use of suitable tools and processes to confidently manage both individual RPA projects and the enterprise-wide automation landscape. It is about efficiently orchestrating the cross-organizational collaboration of processes, people, and bots in line with the automation strategy.

RPA governance can be broken down into three main levels:

  • Strategic 
  • Management  
  • Operational 

3 levels of RPA governance 

Here are the key questions and issues to consider as you develop a framework for RPA governance and as guardrails for a successful automation journey. 

1. Strategic level: How do I plan RPA?

Objectives. What do you want to achieve by using RPA? Do you focus on improved process quality, increased efficiency and associated cost reductions, do you need to reduce FTEs or free up resources for higher value activities, do you want to free your teams from mindless activities or from temporary peak loads? Which departments are you focusing on? The more precisely you describe your business goals, the easier it will be to plan your RPA strategy.

KPIs. What metrics do you use to measure RPA success? These can be business figures such as a 15% reduction in process costs within the first year of operation or break-even analyses that record the point at which the investment in RPA pays off. “Soft” performance indicators such as the increase in employee or customer satisfaction could also be of interest, which you could survey in percentage terms before and after implementation of automation. Last but not least: Who should receive reports on these success figures?

Commitment. How do you get the go-ahead for RPA? Ensure upfront that you win advocates, subject matter experts, and internal sponsors to convincingly position your RPA initiative with company leadership. This means presenting the intended goals and benefits, including the expected return on investment, and providing the most accurate capacity, time, and budget planning possible. 

Tools. Which solution will you use for process automation? Who is involved in the tool selection? Can you evaluate your needs and the appropriate RPA software internally, or do you bring in external consultants who specialize in automation solutions? Check with RPA vendors to see if reference customers are available for discussion. Take advantage of free trials or conduct proof of concepts to compare different solutions in a practical way. In addition to the IT department, be sure to include department heads or key users from the business units in the decision-making process so that they can assess the usability for business users.

Processes. What processes do you want to automate? Select the RPA processes based on the goals and KPIs you have defined. Consider how you will prioritize the implementation of the automations, i.e., which processes you will start with and how you will successively fill the RPA pipeline. You might first automate a smaller “pilot process” that illustrates the function and benefits of RPA to management and the business departments.

2. Management level: How do I guide RPA?

Teams, roles, skills. Which stakeholders are involved in RPA projects? Who owns which roles and what tasks are assigned to them? Who makes which decisions and gives approvals? Who informs whom about what? RPA projects involve numerous roles, responsibilities, work areas and communication channels, ranging from the process owner to the bot developer up to the IT administrator, all of which must be described clearly and distinctly from the outset. 

IT business alignment. Who owns RPA – the business or IT team? RPA projects are often managed and controlled by IT, whereas the bots are developed in different departments or locations via suitable low-code approaches. In this hybrid organization model, the IT department acts as the central “business enabler” while the business department drives innovation “on the edges”. In any case, this constellation requires close and coordinated communication and collaboration between IT and business to jointly achieve the desired goals.

Project management. Which specific stages does a single RPA project run through? Which roles are involved and which activities are carried out in the various project stages? Which quality gates are passed during the course of the project – and who releases the project into its next stage in each case? There is no international standard for the practical implementation of RPA projects, but over the years, the model of the “RPA lifecycle” has become somewhat established, analogous to the software development lifecycle (SDLC) – including the planning, development, testing, and maintenance phases. Some RPA providers deliver their software with additional frameworks and procedures for implementation throughout the RPA lifecycle.

Knowledge management. How do you transfer RPA know-how into your organization? You need a single point of communication or a single source of truth on the subject of process automation. Dedicated subject matter experts can install a central point of contact for RPA or answer typical questions via a knowledge base. 

Examples questions include: 

  • What is RPA, and what can it do? 
  • What are typical RPA use cases, and which tasks can be automated? 
  • Who can help me source processes? 
  • Where can I submit automation initiatives? 
  • Can I get training in the solution? 

If you bundle RPA knowledge centrally and make it accessible to everyone, you will foster an appropriate mindset, increase acceptance, and deepen competencies around process automation – all while anchoring automation sustainably throughout the entire company.

Change management. How will the staff react when the bots come in? The introduction of RPA always involves a cultural change. Get your workforce in all departments on board the automation journey from the start, so they literally “go with the flow” rather than stand up against the digital workforce. 

Engage people in frequent conversation, alleviate any fears of job loss, and highlight the opportunities your teams will gain when RPA bots take over the boring “robo jobs”. Upskill your employees to acquire new capabilities for new areas of work. This can also include process automation itself: Offer RPA training for interested employees and educate them in how to identify automatable processes or how to handle the RPA solution in practice. 

3. Operational level: How do I implement RPA?

Process recording, documentation, evaluation. How do you lay the foundation for your automation landscape? Neat, sufficient and detailed recording of the RPA use cases in the business units and standardized documentation of business processes including all their variants are quintessential as a basis for subsequent automation. In addition, the processes must be evaluated for their automation suitability, and if necessary, optimized before automated.

Infrastructure. Where will the bots work? Before actual process automation implementation takes place, IT must provide a development environment, a test environment and a scalable infrastructure for the operational use of the bots – on dedicated end devices or on virtual machines. Also make sure that all applications involved in RPA processes are provided.

Data quality. How can the bots work? will not open or read analog documents, decipher handwriting, or analyze free-text emails. The input data for RPA must be standardized, structured and digitized so that the bots can read it out and process it further. Ensure they can work with data that derives from web forms, excel sheets or standard emails – that’s why preliminary process optimization is essential.

Security. How do you automate safely? RPA bots work as “hybrids between software and human employees” in live systems with real data. The IT department must grant and manage appropriate access rights for the bots and, if necessary, revoke them again if the need is no longer current. In addition, the RPA solution itself must have an appropriate security concept that protects sensitive data and processes.

Bot development. How do you achieve high quality and speed of bot development? A standardized and simple procedure for bot development should be available, ideally a low-code approach that can also be leveraged among citizen developers in the business units. In addition, it should be possible to create templates from previously developed automations that can be accessed centrally and easily reused in future automation projects. This will significantly speed up the development process over time.

Bot testing. How do you ensure correct bot configuration? For smooth and trouble-free operational use, acceptance tests are required before each bot goes live and should be carried out in a separate test environment – just as with any other mission-critical software before it is transferred to the live systems.

Bot operation. How do you manage and maintain the bots in operation? What happens when bots run into errors and stall? Complete monitoring of ongoing operations includes reliable incident management, i.e. automated alerts, including screenshots and detailed error (cause) descriptions that are sent to system managers.

Change management. How do changes in the applications (that the bots access) or in the processes (that the bots automate) affect ongoing RPA projects? Since the bots operate on the graphical user interface of applications and follow a predefined workflow, they will run into errors when application interfaces or process flows are altered. Here, a well-coordinated collaboration between application or process owners and bot developers is needed. If changes are communicated timely, bot developers can adapt automation workflows accordingly and ensure continued RPA operation.

RPA needs governance if you really want to scale 

RPA governance is a broad field that needs to be fully prepared before you cultivate your automation landscape. There are many issues around people, processes, and technology to clarify and synchronize on, but in the long run, you will reap the rewards of your RPA success for many years to come. 

Learn more about how MuleSoft RPA can help you easily address governance issues – and more!