The Email Management Challenge and Opportunity
Robotic Process Automation (RPA) has quickly become a game-changing, enterprise solution used to automate repetitive tasks. However, the nuances of language make it challenging to accurately analyze, understand and automate processes that contain this unstructured form of data. The ability to automate the management of the valuable data contained in the form of language (emails, forms, chat, social media, surveys, transcripts, contracts, et al.) is possible when RPA is combined with natural language processing (NLP) solutions like the expert.ai Platform.
Email Management is About Language Data
In any enterprise, email represents both a challenge and an opportunity. Consider that:
- Employees spend nearly 1/3 of their time managing email
- 3/4 of a company’s knowledge is stored in employee inboxes
Email, along with other text-based documentation, is among the 80% of business data that is unstructured and difficult to analyze without specific technologies. This important channel is also where customers communicate directly with a business to express their requests for help, their suggestions, as well as their complaints. And herein lies the opportunity—being able to capture this information to quickly solve customer issues and to mine it for insight for upstream functions like decision making and product development.
While this may sound straightforward, the work required to manage incoming email is time consuming and tedious for human operators.
To process even a single request, an agent would have to read the body of the email and its attachments and then route it to the proper agent or service area for resolution. Now, imagine this at the scale of a global enterprise—thousands of emails daily and in multiple languages—and you understand the magnitude of the challenge that such organizations are facing. With more customers using online communications than ever before, you can also understand the magnitude of the opportunity that’s at stake.
So, how are enterprises addressing this challenge?
The Use Case
An expert.ai customer—one of the world’s largest global banks—wanted to create an intelligent alternative to the laborious and time-consuming task we referred to above: email management.
This client is a leading European bank serving customers in more than 40 countries through four service centers around the world. So, in addition to managing an already heavy email load, operators have to do so across multiple languages. As you can imagine, this resulted in a lot of duplication in responding to common and similar requests.
The bank’s toolkit for solving this challenge incorporates a variety of technologies, among them robotic process automation and natural language processing. Let’s look at how these technologies work together.
When an email comes in, a considerable effort is required for a human to read emails, extract relevant information from body text and attachments and categorize them for different purposes. This makes for time-intensive, error-sensitive and therefore, costly work. RPA alone could only follow rudimentary rules for automation; the robots need to understand natural language and be made “smarter” to complete the solution. Expert.ai’s natural language engine ensures that the language-based, unstructured data in email text is understood in a way that enables it to be processed. In other words, it understands the intent of the email and the topics discussed and extracts key elements that will be needed later in the process to help automate responses through RPA.
Together, using RPA, NLP, a translation engine and other tools, the bank created an innovative email management solution: a natural language-based robot.
Today, the solution has been rolled out in one of the bank’s customer service centers, where it processes 100% of emails across six languages, 10,000+ monthly tickets and more than 200 categories, all with 95% accuracy. The plan is to extend the solution to all of the centers where the bank operates.
Lessons Learned from A Successful RPA Deployment
Such a transformational project not only involves technology, but the many people whose everyday work would be impacted by these new processes. The bank worked hard to make sure that employees understood that this initiative was about augmenting capabilities, not replacing jobs.
To do this, they focused on two key aspects that contributed to a successful RPA deployment:
Change the metrics. To measure the return on investment of these initiatives, other approaches look at the reduction of jobs, or FTE savings, to determine an initiative’s success. Instead, our customer wanted to focus on employee engagement, not time savings. To do this, they changed the project metrics and constantly measured employee engagement scores throughout the process. Not surprisingly, they discovered that employees saw the solution as one that relieved them from mundane tasks that they didn’t enjoy.
Build in transparency to gain trust. When implementing any new technology or system, building trust with users takes time. Training the robot on the large number of categories and different languages required additional learning, validation and user acceptance testing, and the bank made sure that all these efforts were visible and transparent throughout the process. By sharing these results, employees were able to see that the solution gained accuracy along the way. Once they saw that the accuracy achieved eventually outpaced human accuracy, this helped build trust in the solution.
