We’ve covered how to use intelligent document processing for email automation in the past on this blog, including this piece on automation for digital mailrooms. But I’m not sure we’ve covered it as thoroughly as Munish Arora did in the most recent episode of Unstructured Unlocked, our new podcast to help Center of Excellence and other stakeholders navigate the world of intelligent automation as it applies to unstructured data.
Arora is Associate Director of Advanced Analytics at Sun Life, the large global financial services and life insurance company. He’s responsible for mining data to build analytic insights and intelligent automation projects that, increasingly, involve unstructured data.
In the podcast, I asked him to describe an intelligent automation project that had a big impact on the organization, forcing it to solve some thorny problems. He described in detail just how long it can take for an insurance company (or any company, really) to deal with even a simple email or phone call – and how intelligent document processing automates customer service requests.
From his explanation, it quickly became clear how intelligent automation can play an important role in contact centers, with respect to more quickly handling email and phone calls. Intelligent automation enables companies to improve customer service while greatly increasing the productivity of their workforce. The contact center is also a multi-faceted use case, where you can apply automation at various points where it brings the most value; it’s not an all-or-nothing proposition.
Listen to the full podcast here: Unstructured Unlocked episode 2 with Munish Arora
Traditional email processing
The process for handling email that Arora described will sound familiar to any process professional. Say a client sends an email saying they want to make a change to an insurance policy, perhaps adding a spouse. Normally, a customer service representative needs to read the email, understand the request, and forward it to another employee who is authorized to make the sort of change the customer is requesting. If the request comes in by phone, the process is much the same, with the telephone rep initiating the change request.
“A typical organization will have multiple processes or multiple tools [involved] before it actually reaches the final point where you would process that request,” Arora said. “It takes time. You would have your request completed after two days, three days or maybe a week.” In other words, it’s not a stellar customer experience.
So, Arora and his team set out to identify which tools or technologies were adding value to the process and where it could be improved. “Do we really need to change the platform or do we need to give automation to those who are listening to phone calls [and reading emails]?” he said. “Can you build something which can read those emails and trigger the next process?”
Automating parts of the process
He came up with a multi-point solution addressing both email and call center agents.
For email, the solution was a text-based intelligent automation model that could read emails, classify them by the type of request, and forward it to an appropriate department accordingly. “You try to automate those things with an unattended bot because emails can come in at any time of day, and you really don’t need a person just to scan it,” he said.
For call center agents, on the other hand, his team built what he called “attended bots” that the agents could use to trigger an automation routine to fulfill caller requests. “The automation will help to file a case, give it to the relevant team to validate the information, and then process it,” he said.
That leaves only the more complex or unusual cases for live agents to manually process.
“We really were able to shorten our time and our clients were happy,” Arora said, noting processes that previously took “many days” now just take one or two.
Process understanding is key to automation
It was immediately clear to me that part of the reason his team was successful in automating this process was they took the time up front to analyze and understand what was involved in it.
That required getting multiple teams involved, including not just the automation team but business users and operations teams, which often include folks versed in Lean and Six Sigma practices. With a full grasp of the various steps in the process, then you can surgically apply automation where it makes sense.
“It was also step by step,” Arora said. Once the team figured out where inefficiencies were, it would build an automation routine in an agile fashion, iteration by iteration. The business would then run each iteration, starting with, say, a bot in attended mode.
Such an approach helps the automation team and COE avoid a common challenge: business groups not wanting to use tools the automation team delivers. As customer service agents incrementally see what the bots can do, “that builds their confidence in us,” he said, and makes them more willing to go on to the next step.
Check out the full Unstructured Unlocked podcast on your favorite platform, including:
Read the full transcript of the podcast here.
