When scaling and streamlining your customer service teams, most company’s first look at how to improve areas such as First Response Time (FRT), First Contact Resolution (FCR) or Average Handle Time (AHT) among others.
But digging deeper into what can work on all three of those areas is the use of Robotic Process Automation (RPA).
RPA customer service can add a rush of speed across many areas of the contact center platform and lift the energy of the team by taking the everyday humdrum away from them.
“If you have any repetitive mundane task that takes more than 30 minutes to complete, it’s a perfect candidate for RPA,” says Ben Segal, Associate Director of Infrastructural Efficiency at Freshly. “Any task that is the same all the time and is done on a consistent basis. That is a candidate for RPA and you should look and see if this may be a viable solution for you. That goes for the agents or for the admins and really for anybody, to be honest.”
Not only that, but there are other uses for RPA that can help in customer service that are “behind the scenes” like gathering data to use in quality assurance.
Freshly started using RPA for the customer service department in 2019 following a 2018 summer customer experience summit and tour of a contact center at another company. Segal says the company invited those who work in CX to “come out and look under the hood.”
Freshly is a food company offering healthy pre-made meals delivered fresh to customers' doors. Freshly is committed to making food that’s delicious, healthy, and convenient.
Freshly started using Playvox in 2019.
According to Segal, the company was using RPAs for handling agent dashboards and getting all the reporting to update in real time.
“I was like, ‘oh, that sounds interesting,” he said “So I went and sat down with the guy who had done that presentation and he explained you could take a free 60-hour class at UI Path and you could even do it online. And at the end, you learn how to build a basic RPA bot.”
Creating An RPA Bot To Do The Grunt Work
Segal and Freshly were really intrigued by how easy it was to create an RPA for an entire company.
“It's low-level engineering effort,” he said. “You don't have to go and bother your developers to hook into APIs to do these calls back and forth. Anyone who takes the class could really learn how to train it to go here, click there, and click here.”
Segal had a CX specialist at Freshly who dabbled in coding video games. He even went so far to create a program at his previous job that would automatically pick the place for the team to go to lunch.
“I liked his mindset,” Segal said. “I mentioned this to him in a one on one meeting and he loved it. He thought it was the coolest thing in the world. He asked to take the classes and I was like, ‘Yeah, go ahead. If you want to do it, go do it. It's free.’ So on his off time, he went for a couple of weeks and took the UI path course and trained himself up on how to make a bot.”
Freshly started to use the RPA internally by pulling general reports from Zendesk and putting them on Google sheets and housing them in a file and running the analytics which as he said was “the Day 1, early level bot.”
When they started using Playvox, they created an RPA that got data from Zendesk and moved it into Playvox to power the agent dashboards. How the Freshly bot does this is it grabs the data from Zendesk and creates a campaign and then the data flows into the corresponding campaign in Playvox.
“We get agent dashboards updated in real time,” Segal said. “If that's done with manual labor, it's a couple of hours of work and it's an admin job that someone has to do once a week and you can only get agents tosee their new information every week. In a world of RPA bots, we were like, ‘oh, we, we could do this every day.’”
So they have it set up to run each night to grab the data and apply it to the agents’ dashboards and the following day, the agents see their results.
“So the bots run and it takes about 20 minutes for the whole process,” Segal said.
RPA Style: The Perfect Double Agent
Not only are the day-to-day tasks of the managers reduced, but surface-level customer service has also been developed by Freshly.
A big task the Freshly customer service team handles is order cancels. Freshly works on a subscription style where meals are delivered weekly to a customer’s door. But Segal said there are times when a customer can skip a week and not get meals delivered or they reduce the weekly quantity.
“A main reason people reach out to us to cancel or skip their weekly order” Segal said. “‘I meant to skip this week. Can you cancel that order? I just got charged. I didn't mean to order.’ It's a weekly subscription. So we're constantly in a world where people are saying, ‘oh, I didn't mean to order this week.’”
So when they examined those scenarios, Segal said when broken down it would take a human agent five and a half minutes of time and 38 clicks of the mouse.
Segal admitted that much of that was due to the way Freshly set up their back end order system and it was never designed to do the number of cancellations.
“It’s not the end of the world, it’s not an hour, it’s five minutes,” Segal said. “But it is 38 clicks of the mouse when you really dig in.”
So they created an RPA bot where an attended agent gives the bot one piece of information like an email address or an order number and the bot would go through and do the 38 clicks to get a result, which would be an email kicked back to the agent saying it was resolved.
“And that might take 30 seconds,” Segal says. “So you go take the email address (or order number), put it into the RPA bot starter. And then let it go and run all the steps that you used to have to do over five minutes. You can even move on, you can take another call to take another email and just know that when the email disappears. It worked.”
Moving Forward In Freshly’s Future RPA Customer Service
As Freshly solidifies its RPA customer service, the next step, Segal says, is moving toward an unattended RPA addition.
“What we see as our end goal, and we were probably weeks away at this point, is exposing that starting point to the end user,” he said. “Letting the customer kickoff the RPA via chatbot.”
What it means is a customer who wants to cancel an order could respond to the automated live chat by putting in their email address and the bot would find the information for that customer. The bot would reply with several order options for the customer to choose from.
Following the customer’s choice, the bot would send a message confirming that cancellation.
“They're truly self-serving,” Segal said. “Now, the customer is thinking ‘I don't need to talk to a human. And I feel like, yes, I talked to a robot, but it was an enjoyable experience where it was just that easy. It just worked and I refunded myself.’”
The goals in that form of RPA customer service leave the consumer with an easy experience as Segal says “they didn’t have to deal with pushback,” and it frees the human agent to do more involved tasks.
In summary, Segal sees RPA as a low tech answer for any small to medium business to take advantage of. At its core, if, say a CX team needs a bot created and configured, but the engineering team sees it as a low priority, they may not be able to get to it.
However, the relative ease in creating an RPA bot can be attained by virtually anyone and can be utilized by the company.
“Say they're not going to spend the expensive engineering resources to fix the problem that we need solved when they have a backlog of a ton of things that are of higher priority,” Segal said. “That is where RPA solves a problem. If CX could find a way to build a simple RPA bot to do some of those actions that are current pain points for CX and I don't need to bother an engineer, it’s a true win win.”
Are you utilizing RPA in your customer service operations? Let us know in the comments!