Desktop version

Home arrow Marketing

  • Increase font
  • Decrease font


<<   CONTENTS   >>

Robotic Process Automation

As the transaction volumes in IT ecosystems increase, automation is being applied to reduce manual work tasks that have high cost, long lead times, and lower quality. RPA automatically accesses applications to create, read, update, and delete (CRUD) metadata and data. Sometimes an automation initiative is large in scale and creates entirely new applications or features

TABLE 8.3

Workflow Management

Capability

Method

Comment

Manual processes and systems

Simple functions (e.g., calculating and checking text)

Ad hoc design based on local content

Virtual versions of manual operations and systems

Process monitoring and reporting

Standardized process structure and definition allowing systematic improvements to the process over time

Virtual versions of the process with modeling and analytical capability

Process optimization based on current or predicted data

Rapid ability to analyze a process to create alternative designs and optimize the best design

Adaptive processes enabling simple responses to changes in inputs

Process reconfiguration due to changes in system status caused by variation of input levels and external influences.

The system changes the process design based on rules and logic programmed by its users to optimize certain critical outputs

Advanced adaptive systems that reconfigure a system dynamically to optimize several outputs

Optimization algorithms to ensure goal alignment and convergence

Characterized by an ability to reconfigure, add, delete, and modify processes dynamically

and functions having higher performance. In other situations, islands of automation are applied to a sequence of process steps based on business rules and a macro algorithm that improves process efficiency and quality. These focused applications are called RPA projects. RPA applies software tools to complete tasks formerly performed by humans. RPA differs from an enterprise application in that the focus is on lower-level manual work activities that interface with larger IT applications. A common application is the use of macros to replicate work activities step by step.

These processes must be rule-based in that software code can be built with defined inputs and outputs. These are modeled using if-then statements, e.g., if conditions A and В exist, then update this status. This requires the inputs to be structured so they can be read by the system. The format is structured as numbers, optical character recognition, or text searches using rules (e.g., frequency counting). Processes that rely on unstructured data that cannot be reliably read are not good candidates for

RPA. A manual process should also be repetitive, and its work activities should follow a sequential pattern that repeats. The process must also be mature and stable with large volumes or batch sizes. Finally, the manual activities should have a high cost or quality levels should be low so the return on investment for RPA is high. Examples include filling metadata fields with customer data based on a DUNS customer number. Once the RPA robot (bot) reads the DUNS Ultimate Customer ID (UCID), additional metadata from other systems can be collected by the bot to build a customer profile. Other examples include inserting delivery addresses in an e-mail or creating an e-mail targeted to specific persona and use cases based on business rules.

Table 8.4 shows three steps. These are creating the RPA foundation, transforming the process for higher productivity, and sustaining performance improvement. The team begins by confirming feasibility (i.e., the process is mature, stable, rule-based, and will exist for the foreseeable future). A formal project charter is created that defines project scope, deliverables, and their schedule. The next step is to walk the process by capturing screen shots and recording the sequence of operations needed to complete the work task. The current-state process is documented to identify automation opportunities based on known RPA success criteria. The team also captures the business rules that govern the creation of the work product of the process. There may be variations of the types of work done by the process based on user persona and their use cases. Other relevant actions at this RPA step include confirming baseline metrics. The focus is on the confirmation of operational and work task rules. The automation is tested under controlled conditions and, once approved and shown to be effective, it is turned over to the process owner. At the end of the project, process documentation is updated and roles and responsibilities are modified to reflect the higher automation.

 
<<   CONTENTS   >>

Related topics