So, your organisation has decided to implement an RPA robot. After careful consideration, it has been decided to automate a process that contains a high level of predictive, repetitive activity with structured data that traditionally has been handled by an employee or a group of employees.
All the prerequisites have been fulfilled as to the formal parts: The project has been sponsored by the COO, CTO or CEO. A steering group has been appointed. The business side have been appointed as the future owners of the robot.
The project can now kick off. We use an agile project methodology as we deal with the development of software. Implementing an RPA robot can be divided into four parts. Each having its own distinctive start and end.
Depending on the organisational culture and its bureaucracy, experience tells us that the first two stages may take up to 70% of the time of the project. This is the pre-work, dealing with the business side, before starting to design the robot, testing it and implementing it.
The four phases and activities within
In the assess phase we investigate the process (or processes) that could be automated. Questions that we want answers to are:
- Is the process repeated several times a day?
- With a high degree of predictable outcome?
- Does it deal with structured data?
- Is the process cross organisational or is it contained within a department?
We also evaluate key criteria such as KPIs and critical success factors for the RPA implementation. These need to be agreed and set before the implementation.
The stage ends with an assessment report that outlines the RPA project in more detail and the feasibility of it.
The second phase, the approval stage, is labour intensive as it is during this stage an agreement about the pilot process to be automated is reached. When a decision about which process to automate first is reached, it is followed by a thorough investigation and documentation of the process.
After documenting the process in detail, we would also design the future process – the one with the robot. It will look different as some of the human interaction and activities are now performed by the robot.
Placing the current process chart next to the new process chart will visualise how the introduction of the robot will alter the process.
This stage ends with a Business case being presented to the steering committee and the sponsor of the project. It contains the expected cost and the Return on Investment (ROI).
At the design stage we would look at the process and start looking at which software vendor best fulfills the criteria outlined in the business case. Different software have different advantages which might and these might be better suited to the organisations needs and future demands. When that choice has been made we would acquire the license from the preferred vendor.
During this phase we would design the exact process and have it signed off by the business. It is then a question of designing the robot.
This stage ends with testing the robot. The robot should first be aimed for quick wins – activities in processes that do not add value but take time and effort. With agile iterations, the robot will be programmed to increase the level of automation.
This is an iterative process and several iterations are performed as the robot is fine tuned. When the robot mimics the users behaviour to a very high degree and when all exceptions have been programmed it can be ready for release.
Implementing the robot is the exciting part. The robot is now released into a real working environment, mimicking an employee’s behaviour. When installed, it is up to the business side to monitor the robot and handle any exceptions. If there are changes in the process, the robot may have to be reprogrammed and it is the business side’s responsibility to either do it themselves (if they have the programming knowledge) or bring it to IT’s attention.
After having been in operation it is time to measure how productive it is and what the business impact is. Is the process faster? Do we save time? Do we save resources? Does it increase revenue?.
It might then be time to look at your next automation project.
Richard Ahl PhD is Program Manager at Clancarty Consulting Ltd based in London