#1 - Achieving digital transformation through RPA and process mining
Bi-weekly newsletter series highlighting the stories and experiences of technologists, startups, and entrepreneurs who change the future of work through RPA (Robotic-Process-Automation), Intelligent Automation, Artificial Intelligence (AI), Process Mining, and broader automation technologies.
We will cover how a technology area within the Future of Work space impacted and changed our lives. For the first one, let's start with RPA (Robotic-Process-Automation) and Process Mining
Understanding what you will change is most important to achieve a long-lasting and successful robotic process automation transformation. There are three pillars that will be most impacted by the change: people, process and digital workers (also referred to as robots). The interaction of these three pillars executes workflows and tasks, and if integrated cohesively, determines the success of an enterprise wide digital transformation.
Robots are not coming to replace us, they are coming to take over the repetitive, mundane and monotonous tasks that we’ve never been fond of. They are here to transform the work we do by allowing us to focus on innovation and impactful work. RPA ties decisions and actions together. It is the skeletal structure of a digital process that carries information from point A to point B. However, the decision-making capability to understand and decide what comes next will be fueled by RPA’s integration with AI.
“ From a strategic standpoint, success measures for automating, optimizing and redesigning work should not be solely centered around metrics like decreasing fully loaded costs or FTE reduction, but should put the people at the center”
We are seeing software vendors adopt vertical technology capabilities and offer a wide range of capabilities to address the three pillars mentioned above. These include powerhouses like UiPath, which recently went public, Microsoft’s Softomotive acquisition, and Celonis, which recently became a unicorn with a $1 billion Series D round. RPA firms call it “intelligent automation,” whereas Celonis targets the execution management system. Both are aiming to be a one-stop shop for all things related to process. We have seen investments in various product categories for each stage in the intelligent automation journey. Process and task mining for process discovery, centralized business process repositories for CoEs, executives to manage the pipeline and measure cost versus benefit, and artificial intelligence solutions for intelligent document processing. For your transformation journey to be successful, you need to develop a deep understanding of your goals, people and the process.
Define goals and measurements of success
From a strategic standpoint, success measures for automating, optimizing and redesigning work should not be solely centered around metrics like decreasing fully loaded costs or FTE reduction, but should put the people at the center. To measure improved customer and employee experiences, give special attention to metrics like decreases in throughput time or rework rate, identify vendors that deliver late, and find missed invoice payments or determine loan requests from individuals that are more likely to be paid back late. These provide more targeted success measures for specific business units.
The returns realized with an automation program are not limited to metrics like time or cost savings. The overall performance of an automation program can be more thoroughly measured with the sum of successes of the improved CX/EX metrics in different business units. For each business process you will be redesigning, optimizing or automating, set a definitive problem statement and try to find the right solution to solve it. Do not try to fit predetermined solutions into the problems. Start with the problem and goal first.
Understand the people first
To accomplish enterprise digital transformation via RPA, executives should put people at the heart of their program. Understanding the skill sets and talents of the workforce within the company can yield better knowledge of how well each employee can contribute to the automation economy within the organization. A workforce that is continuously retrained and upskilled learns how to automate and flexibly complete tasks together with robots and is better equipped to achieve transformation at scale. The time relieved through automation can be reallocated for this purpose. This leads to the importance of having a strategic citizen development approach in your organization that will increase the adoption of automation. Citizen developers are employees who use low-code or no-code platforms to rapidly create automation solutions for themselves and for their team’s needs. Do not underestimate the power of network effects within an organization. A successful robot that brings value to a business unit with measurable outputs and a well-prepared business case will demonstrate to other business units what they could be saving or making as well. An automated process that brings value triggers a rush for robots similar to the gold rush — all business units will pressure the RPA CoE to analyze their processes as well. Network effects create an RPA community/advocates within the organization. A culture that facilitates knowledge transfer between solution architects, leads, RPA developers and business analysts will lead to higher code quality, easier maintenance and greater productivity. If the community thrives in your organization, they will be better at communicating automation prerequisites, discovering process bottlenecks and setting the expectations of the business.
Data science, RPA and process excellence CoEs can complement each other and share important skills that can fill gaps to achieve thorough process transformation. Data science can feed the knowledge to build ML algorithms for the RPA teams, and process excellence teams can help target areas for automation in a business process with a holistic and data-backed view. Setting clear objectives, responsibilities, and expectations and defining a culture of transparency will bring more intelligent automation.
Understand the process
Hidden and uncaptured manual decision steps, memorized actions, undocumented process steps and inaccurate KPIs are the norm in every organization. Data is the new oil — collect as much of it as possible about your processes before diving into optimization, redesign and automation. Your goal is not to automate everything, but automate smart. We have seen processes that are not well understood or analyzed and are automated with limited knowledge — these later increase deadweight cost and miss ROI targets, as defined in the business cases, and can be costly down the line. First, look for process bottlenecks, areas of inefficiencies, or actions with high rework rate or throughput time based on the goals you defined. Emphasize understanding the process and strive to obtain data-backed KPIs. There are two technological capabilities that power process transparency and process optimization, redesign and automation.
Process mining (accesses event logs of multiple IT systems to build a system story) — By mapping digital footprints and pathways, process mining reveals the active steps of a given process. The derived data makes it possible to transform complex processes into easy-to-absorb visualizations, providing the foundation to optimize operations companywide. Process mining provides many benefits, such as discovering pain points, streamlining efficiencies and cutting operational costs. It can be used in any process that has a digital footprint.
Task mining (leverages computer vision to recognize user actions occurring on the screen and application windows to build a user story) — Contact centers are a good example of using task mining to capture process steps and relevant KPIs. Decreasing the average handling time in a phone call is one of the most important metrics that demonstrate customer and employee experience. Capturing different process flows can yield both process-related metrics and maps.
Process and task-mining solutions can automatically transfer the process knowledge and well-defined process metrics into an actionable process map and PDDs/SDDs to help build the automation. UiPath, for example, can output automation files based on the data captured, which accelerates implementation timelines.
Both task and process mining are powerful AI tools that can help you capture and manage your processes. However, pay special attention to the requirements of the software so they can work effectively. For example, process mining projects can be challenging if the data is not cleaned and available in the IT systems. Getting data access to certain databases can be challenging at times. Task-mining projects also need special attention — since task-mining software records the user clicks on the screen, data security (with non-air gapped cloud vs. on-prem) and PII masking will be key for your sensitive processes
Automate to innovate
Before jumping into automation, emphasize achieving the goal set in the beginning by first improving or optimizing the process. We found that some benefits can be realized by applying small process changes instead of directly jumping into the RPA implementation. Once the process’ future state is well defined, there are a few important considerations for effective solutioning:
Intelligent document processing: The world’s most common manual processing revolves around extracting information from scanned papers, handwritten texts, images or forms. Be it for invoices, background checks, purchase-to-pay or revenue recognition processes, IDP alone impacts many of the businesses in HR, finance, IT and supply chain. By cutting manual work, minimizing mistakes and decreasing throughput time that results in faster response times and better customer service/experience, IDP has a very important role to play in your automation journey.
Start with a small scope of documents that has the highest volume and measurable business impact. Once those are digitized and extracted with high accuracy, then increase the scope. It’s important to know that you will need a partner that allows the data classification and extraction models to be retrained and provides good customer support to cut your costs during implementation. This space has been growing rapidly as companies like Base64.ai attempt to solve these problems with a low-code/no-code approach and ready-to-use integration for vendors. Pre-defined models, human-in-the-loop, single-API integrations and the ability to continuously re-train models are some of the capabilities that need to be considered.
Attended automation: Interactive forms and low-code/no-code application builders to replace legacy apps will be key to get your employees to work together with the robots. We have seen custom applications and forms being built to provide better customer/employee experience to the users. UiPath, Microsoft and NICE have product lines built just for that. Robots work in tandem with the employee, pulling data from systems preemptively and collecting inputs from users and passing that data to different IT systems.
API integration and UI-based automation: UiPath’s acquisition of Cloud Elements demonstrates that RPA is not only automating legacy systems and webpages through UI interactions, but will bring ease of integration through the use of a broad range of APIs, too. The UI is prone to errors due to rapid changes, while API connectivity brings reliability and less maintenance. Consider API automation in your workflows before diving into UI elements if the UI isn’t stable enough. This will also allow your solutions to access systems that are both modern and legacy with ease
Measuring operational KPIs of robots: Continuously measure robot performance by analyzing logs and checking the areas/activities that produce the most system exceptions. Address these actions/steps by introducing more automation or process improvement. If built robustly, an automation solution can be iterated over many times. KPIs like robot average handling time, cost per transaction item in the queue, license utilization and number of system exceptions are important metrics to optimize your RPA operations and decrease costs.
The road to the fully automated enterprise
As RPA vendors become a one-stop shop for automation transformation initiatives, their products will grow to address every single area within the automation lifecycle. The industry is in the early stages and is seeing rapid growth with IPOs, acquisitions and funding rounds, and more startups are rising around them. Adoption of RPA and process mining in your organization will define the operational excellence of your firm. If you are behind in this race, just think of how your enterprise can continue to compete with fully digital peers. Your organization won’t want to be in the back of this race.
