Implementing basic RPA is remarkably simple. The task is performed by a human as usual and "recorded" for the software. The resulting script is fine-tuned to ensure all potential task variations are accounted for. The resulting robot is tested to ensure it works.

A strategy that incorporates RPA as part of a more sophisticated automation capability — one that involves AI, data capture, business rules or workflow management — is likely to require additional planning and scrutiny of the overall process, including all inputs and outputs.

It helps to first identify the tasks that are most appropriate for automation and lead to improved ROI. This will help determine whether it’s wise to consider an enhanced solution, or continue to perform the task manually, instead of implementing basic RPA.

• Tasks that are simple, consistent and repeatable.
• Repetitive low-skill tasks that are prone to human error.
• Existing or planned processes where automation can improve productivity and efficiency.
• Tasks that can be executed quickly to improve customer and worker experiences.

Some tasks may meet many of these criteria but would still not be suitable for basic RPA. For example, processing paper invoices may meet every criterion, but require additional data capture capabilities or a redesign of the process to make it fully digital.