Interactive voice response (IVR) systems are often the first point of contact for customers as part of an enterprise’s customer service system. Unfortunately, IVRs often fall short on the promise of providing ‘self-service’ solutions for customers as IVRs are only capable of connecting to systems that are accessible via API, which even when available can be expensive to implement.
With Robotic Process Automation (RPA) and UiPath Enterprise RPA platform, IVRs can be connected to any system, drastically expanding IVR fulfillment capabilities to gather information and perform actions across a plethora of systems from on-prem mainframes to virtualized systems to cloud-based web services. Increasing the success rate of customer requests being fulfilled by the self-service IVR frees your contact center agents to work on more complex customer issues.
The results? Improved customer and employee experiences, enhanced accuracy, reduced average handling time (AHT), and acceleration of digital transformation initiatives, resulting in a rapid return on investment (ROI).
This article will show how to integrate UiPath automation with Amazon Connect to implement an autonomous, self-service contact center solution in your enterprise. Due to the 2019 novel coronavirus (COVID-19) crisis, customers are reporting an increased need for outbound calling automation so I've also included an example in this post.
Fulfilling IVR requests with UiPath automation
Before we go into the details, consider a typical scenario of a customer contacting the customer service department at for a telco company:
Aileen is a customer of Meeple Telco and wants to know the current balance for her mobile phone bill. She makes a voice call to her Meeple Telco and the IVR validates her identity by having Aileen provide the last four digits of her social security number. Aileen is prompted with a menu of options. She chooses to get her current bill balance and the IVR flow fulfills this request and reads it back to Aileen using text to speech. Aileen is thrilled that she was able to get an answer to her question in under 30 seconds and without having to talk to anyone!
This scenario may sound simple, but the devil is always in the details. How exactly does the IVR fulfill this request? A developer could potentially connect to the back-end systems with APIs, but what about when APIs aren’t available? What if there aren’t resources available to build these bridges across systems?
This is where the power of RPA comes into play.
Figure 1 Overview of how UiPath can be leveraged for fulfillment in Amazon Connect
Since the back-end system is not accessible through APIs, we call UiPath Orchestrator to run an unattended automation that accesses these systems for us. We pass the caller’s phone number as an input to start a job on UiPath Orchestrator.
An unattended robot then picks up the job and runs an automation to gather Aileen's balance from a backend mainframe system. The job returns the balance as an output. Our solution centers around Amazon Connect’s ability to call external systems via Amazon Web Services (AWS) Lambda functions. The IVR contact flow is simple:
1. Start a UiPath Orchestrator job with inputs (StartJob AWS Lambda function)
2. Check every three seconds to see if the job completed (QueryJob AWS Lambda function)
3. When the job completes, extract the output and play the result to the user
Figure 2 An example Amazon Connect Contact Flow invoking UiPath automation to return information from a back-end system to the customer.
Since Amazon Connect is an omni-channel solution, the same flow will be used to support customers contacting us either over voice or through chat. Here’s a quick video showing the flow in action for a customer reaching out to the contact center through chat:
While contact centers traditionally focus on reacting to incoming customer requests, there are many scenarios where proactively calling customers is required. Due to the COVID-19 pandemic, we’ve seen an increased demand for automated outbound calling scenarios where companies need to reach out to a customer and have them answer questions. One example of this has been the flood of citizens requesting to be placed on unemployment. State governments can’t scale their call centers to achieve the thousands of hours it would take for humans to make these calls and process the resulting data.
Our solution uses an automation to pull data from back-end systems, enumerating the applicants that have missing data. RPA instructs the Amazon Connect IVR to call the applicant when missing data is identified. During the call, the IVR asks the applicant to provide the missing data in a fully self-service manner, using either simple numeric inputs on the keypad (i.e. age or gender) or using LEX to process natural language responses (i.e. addresses or employer name).
Below is a simplified version of this flow where RPA processes a set of records in a Microsoft Excel spreadsheet and instigates a call to the applicant through Amazon Connect to provide their social security number.
Setting up the automated contact center solution
If you are interested in trying out this solution yourself, we’ve put together a starter kit with everything you need. Please visit the UiPath\AmazonConnect GitHub repository for more setup instructions and all the necessary code samples. You can also contact our sales team directly.
Justin Marks is a principal product manager at UiPath.