How to integrate RPA into automation toolbox of your company
Robotic Process Automation (RPA) has in recent years became known as technology enabling fast reduction of repetitive, manual tasks. But how to take RPA from hype to a corporate solution?
Automation needs might sometimes emerge suddenly without months of planning and prioritization. Like few months back I was contacted by Claims Unit Director “Hi Asko, I have something to tell you, and you will probably not like it…”.
He told that they need to send a total of 5000 letters to customers for taxation purposes within two weeks. Letters had previously been sent to customers that had received compensation automatically from an old system that had now been closed down, but the functionality had not been implemented into the new system.
The information needed into letters required complicated retrieval of data and would take months if done manually. Within same day we gathered the claims experts and RPA professionals into common one-hour Teams meeting and concluded that the task was doable with software robotics and most likely also with the given timeline.
At the same meeting it was decided that the robot needs not only to compose the letter, create the pdf, and send it to mass printing, but also to archive it with correct metadata to be available for Claims Handlers possibly answering to customer questions.
Within a week not only the robot was build, but all the letters were send to customers in legally binding timeline and problem solved also for the future as the robot would do the task also years to come.
What is RPA?
Robotic Process Automation aka RPA is a software robot that is taught to use existing applications with their own user IDs and passwords just like human workers.
The robot does not learn anything by itself, but each stage is taught in detail. By default, using a software robot does not require programming language skills or require changes in the target systems, but technical background and understanding company`s IT application landscape can help to build robust robots.
If`s organizational and technical RPA setup
In If the whole RPA team has been recruited internally from different professionals in different roles within the company. Developers have long background with working with and developing If IT systems.
Knowing the applications, way of working and the actual people around the processes has proven to be beneficial in not just educating people about RPA possibilities but also in building maintainable RPA processes.
We started building our RPA capability systematically from 2018. From the beginning goal was to create a centralized Center of excellence that could efficiently support all the Nordic countries and business areas with their automation needs.
The RPA technology chosen was Blue Prism which at the time and still is commonly used in the banking and insurance companies. RPA technology market is very competed and other solutions such as UI path, Automation Anywhere and smaller players or even open-source RPA options can do the tricks that you expect from RPA software.
With a continuous development with our partner, we have been freed for a quite a while from a typical pain of managing needed RPA licenses and efficient utilization of them. If´s virtual machines where our robots live in Azure cloud are managed by our partner.
We do not buy separate licenses, but there is unlimited RPA capacity in the cloud which is paid by the minutes the robots actually work. Prioritizing and orchestrating the work of the robots is done based on given SLA`s automatically by tool called RPA Supervisor which usage is included in the minute-based pricing.
Microsoft continues also developing their RPA capabilities. In if we use Power Automate to complement Blue Prism capabilities especially in so called attended RPA use cases. Most of our automations run unattended in the background triggered by a task coming to a work queue to email box etc. or run based on certain schedule. In attended RPA task is triggered typically by a human user to carry out a task or a process on behalf of to human often immediately sometimes during a customer contact.
In If RPA setup MS Power Automate is used to “sniff” these types of tasks from users and then trigger the Blue Prism based automation in the cloud to carry out the task. Therefore, we do not need to use Blue Prism capacity and minutes and to stay alert but are able to provide Attended automations that are quick to run without delay as fast as the target systems the robot is using are able to respond.
Combining RPA capabilities with AI
RPA is not considered as Artificial Intelligence as it is completely rule based and does not learn from the data that is processing. Artificial Intelligence on the other hand can complement the abilities of RPA which needs structured input what especially machine learning (ML) can help to build. ML can help to transform unstructured data such as text, images and voice in a more structured form that can work as input for RPA robot.
If has used AI based chatbots to serve internal and external customers from 2017. Current common platform for building AI based conversational automation is provided by Boost.ai.
Customers can for example ask from customer service chatbot Emma about how to get a corrected invoice (significant share of customer service contacts are about invoicing). After identifying the customer need Emma can direct the task requested by the customer to its colleague RPA robot which goes through the insurance changes of the customer, updates the invoicing and informs the customer.
So how about that automation toolbox?
Like any other company If has ambitious goals towards digitalisation and automation. Aligning with integrations, API strategy, renewing of core systems and other initiatives is important. In If we have managed to do this by close co-operation with our architects and agile operating model.
By creating a one-stop shop together with If tech community we can create significant business value with robotics. Business only needs to show their problem (in practice, not with slides) and right tool to solve the automation need is typically decided in on hour Teams meeting.
Transparency creates efficiency
So far Robotic Process Automation has been used in dozens of business processes in If and RPA robots have completed millions of tasks in automations in all Nordic countries within various use cases typically found in customer centers and in claims processes.
We have built a common dashboard called Nordic RPA report where anyone can follow what robots all over If are doing and not doing. Meaning all the exceptions that robots have not been able to handle are visualized in the report and can be used to further develop the automation. Often this sort of information would not be visible if the tasks would be carried out by human.
Regarding integrating RPA as a natural part of your automation toolbox. It all starts from your company´s desire and target to improve processes and automate. If that type or sense of urgency does not exist of course also selling of RPA services internally takes a bit more time.
Whether you have top management driven automation agenda or more of a bottom-up approach, I recommend building common and agile ways for solving the automation method together with rest of the tech organization.
Therefore, you have a good chance of delivering business value quickly and also good possibilities of reaching the “final” automation solution when it is visible for whole company what you are using RPA for.
Read more: https://www.linkedin.com/pulse/combine-emotional-artificial-intelligence-wisely-asko-mustonen/
