Robotic Process Automation (RPA) is changing the way enterprises manage their front & back office operations by combining the power of Optical Character Recognition (OCR), Natural Language Processing (NLP), and Machine Learning (ML) to make processes intelligent and efficient. Benefits of RPA include improved accuracy and speed as well as reduced error rates resulting in significant savings.
Accounts payable & receivable, claim processing, policy issuance/claims, mortgage document processing etc. are some of the common paperwork intensive processes that can be automated with a high ROI. In most scenarios, these processes require extracting data from structured and unstructured documents using OCR and other techniques.
OCR technologies have made tremendous strides in understanding and extracting relevant information from documents using techniques like auto-correction, fuzzy logic, rule-based conversions, and more. However, we haven’t quite reached the point where we can be completely dependent on these intelligent automation technologies. Sometimes a simple mistake to identify a decimal place can put the organization at financial risk.
In general, the processes above are executed under a controlled environment where the automation bot is assigned to perform the repetitive & time-consuming work like extracting data and applying logic to auto-correct and hand it over to a human counterpart to validate, and correct, if required.
In these scenarios, the individual who manages approvals (i.e. the process owner) needs to be attentive to prompts while the automation bot is executing, as the bot needs approval from a user before further processing can occur. This inter-dependency results in wasted time where the bot may be waiting for user input or the user may be waiting for the next anomaly to review.
The below graphic highlights a typical accounts payable document process flow.
Note: This is a common flow regardless of toolset (UiPath, AutomationAnywhere, Blue Prism, etc.)
As you can see, in this process, both the bot & user needs to be well synchronized to execute the process successfully. This is where actionable messages can really help.
Actionable messages* are special email messages formatted in the form of “Adoptive Cards*” containing the data that the user needs to take quick action on – within Microsoft Outlook or Microsoft Teams.
When an automation bot encounters a scenario where a user needs to confirm the validity of data, it can send an actionable message to the process owner with the information captured thus far. The recipient can then review the data within their outlook email client from anywhere and at any time to change the data, if necessary, and send the approved data back to the bot via email. The bot then reads the properly formatted data or approval, updates the necessary backend systems and proceeds with completion of the automation task.
These actionable messages remove the synchronous dependency during the process which improves process efficiency.
Let’s now change the process flow to take advantage of actionable messages.
As you can see in this approach, when the user can respond to actionable messages immediately – even when on the go, the Bot can then process requests faster with better accuracy.
Exploring how you can modernize your back or front office experience through the power of process automation? Learn more about Robotic Process Automation and other intelligent automation offerings. Contact us for a complimentary consultation or POC.