I have a few bots, what now?

I have a few bots, what now?

Congratulations!

You just went live with your first RPA implementation. Unless it was part of an overall strategy, you might be wondering what to do next.

Let’s quickly examine why many RPA efforts fail or never move beyond the science project phase:

  • Less than optimal processes were selected for RPA implementation. Processes need to be highly repetitive and stable, using structured data, where the decisions required are rules-based.
  • An overall strategy with C-suite support was not established.
  • Comprehensive change management across multiple teams was not established.
  • The company’s current culture resists change/adoption, leading to delayed timelines and project cost overruns.
  • Competency within business SME community to identify suitable processes for automation was not established
  • RPA development competency to create and manage the type or complexity of robots needed was not established.
  • Appropriate guidelines driven by an RPA Center of Excellence (COE) that set the standards for process selection (including an ROI calculation), development standards, quality assurance, maintenance, and value reporting were not established.  Summary: lack of appropriate Governance.

That one-word “Governance”, and all that it implies is one of your most essential considerations when plotting your course for RPA expansion. You need to consider and plan for each phase of RPA development and deployment and how you will govern each, to create a self-sustaining RPA assembly line with the necessary controls and reporting.

A first consideration should be, how is your initial implementation measured, and by those measures, is it successful? Do these measures include a total cost of ownership (TCO) calculation, and does that calculation scale and align with management expectations? Necessary measures as you expand your RPA footprint as evidence of a great ROI are always compelling.

How do I determine which processes to automate?

Ideally, you are in an industry that is ripe with opportunity. Even so, you still have many options. You identify suitable process candidates through task capture or mining and process mining. Depending on your circumstances and level of maturity, one approach may be more suited for you. Let’s look at each.

Task Capture

Task Capture comes along as you move through a work process you’d like to automate, taking screenshots and gathering data for each step. Then it pulls everything together into a document—ready for dev teams to start automating. Task Capture can automatically generate a process map from application screenshots and window names, along with titles and descriptions of each step.

Task Mining

Task Mining is used to identify and aggregate process workflows and apply AI to map tasks to automation opportunities. Leveraging centralized control from an admin portal, you can securely capture and aggregate employees’ detailed workflow data—including steps and execution time—without interrupting their work. Most tools then use advanced machine learning models to tease out the most frequent task patterns from the data. Then they identify repetitive activities that could be automated. Machine learning models score automation candidates for both potential savings and ease of automation. Task Mining collects data from many employees doing the same task—then uses AI to aggregate it into a process map along with the drill-down details. Task Mining captures only whitelisted applications. It anonymizes and aggregates individual data. Data upload and transfer are encrypted and secured.

Process Mining

Process mining focuses on gathering enterprise data (referred to as event logs) from corporate IT systems for further analysis. Based on event logs, process mining software extracts existing data about what happened in a process and when it happened. Then, the software algorithms translate the data into comprehensive language and turn logs into a visual workflow. Looking at the actual end-to-end process, you can spot any deviation or bottleneck.

Remember, all the while, you will be improving the quality and the quantity of data to start leveraging other AI solutions. Did you include that in your ROI calculation?

There are many steps to achieve success, but they won’t always follow the same path.

Do I really need a COE?

As you gain experience building robots, the tasks and processes you automate will become more complex. For example, Robots requiring human intervention when the robot hits a limiting decision point or data is missing or irrelevant will be developed. As you build more sophisticated robots, you must include rules for how and who will engage them…..that word “governance” again.

Suppose your implementation is going to grow beyond a science project. In that case, you must consider establishing a COE or making sure the responsibilities of a COE are encompassed within an existing organization.

A COE’s duties may vary by organization but simply put it owns the entire life-cycle of your business process automation activities.

The COE maintains standards for business process selection, automation, and monitoring. The COE is also responsible for measuring and reporting on the overall success of the program.

Inputs change, and robots break. All tools provide monitoring management capabilities to help manage the performance and health of your environment. These tools can continuously provide information to improve process efficiencies back into the organization.  Having the COE perform  this monitoring and reporting is vital.

Keep in mind that you will eventually reach a point of diminishing return for task automation from an ROI perspective, and your resources will migrate towards robot monitoring and maintenance and reporting of the process performance impacts to senior management. Growing your COE in parallel with your business process automation development is undoubtedly recommended.

Are you ready to automate a few more processes?

Most likely, you are. But before moving ahead, take the time to reflect on the points below.

  • With some experience, are you still comfortable with your RPA tool selection?  Are you ready to make a more significant investment in hindsight?
  • Do you have a plan for C-suite understanding and buy-in?  Have you identified a business champion?
  • Have you gauged the company’s willingness and ability to change and scheduled your change management activities accordingly?
  • Do you have a multi-year budgets planned or only approved for the current year?
  • Do you need to procure more licenses?
  • Are you planning for local instance or CLOUD hosting? (Due to how fast these applications are evolving, I recommend CLOUD). Do you have the appropriate production controls covering change management, backup, and DR?
  • Do you have a roadmap of target processes that adhere strongly to your criteria for RPA?
  • How will you create a pipeline of new processes for automation consideration?
  • Do you have an approved calculation for measuring RPA ROI for automated processes?
  • Are you prepared to continually measure and report business improvements?
  • Have you established a COE with cross-functional membership and defined their scope and authority?

This list is by no means exhaustive.  Like every business change program, it requires “planning the work and working the plan” with a particular focus on change management.

For more information, please visit Votum’s blog page at https://votumtg.com/blog/.