The public engages with federal agencies across multiple channels today: social, websites, chatbots, text messages, in person and over the phone. Despite these options, most citizens still overwhelmingly prefer to pick up the phone.
“Even with all the digital tools out there, over 60% of citizens are still making a phone call,” said Andy Beamon, senior vice president of strategic innovation for citizen services at Maximus. “With over 30,000 customer engagement center agents deployed at public sector agencies, we find that interactive voice response (IVR) is really the crossroads of where the experience breaks down.”
IVR is the ideal point to focus on when seeking to improve customer experience, particularly when it comes to implementing artificial intelligence, Beamon said. But implementing this type of technology can be a daunting proposition for a contact center management team that must manage human resources as well as investments in existing technologies, from web apps and chatbots to website and onsite tools, he said.
We asked Beamon to share a mini roadmap for improving IVR using robotic process automation and AI to ultimately enhance customer experience.
Step 1: Don’t start with cost — focus on cost avoidance
People get worried about the cost, he said. Instead, the true cost is the opportunity cost, the cost avoidance that an agency can achieve. If an organization takes that approach, any initiative becomes about recouping dollars that can then be used to provide more services to citizens, Beamon said.
“If we’re looking at it as purely additional cost, we’re thinking about it the wrong way,” he said. “We really should be looking at it as the longer we wait, the less efficient we are and the less we can do to provide good CX to our customers.”
Step 2: Start with a pilot
Expanding the use of AI in a customer engagement center does not have to be an all-or-nothing proposition. In fact, Beamon suggests using a pilot project that can be used to scale out a broader effort.
“If you want to stop, it’s easy — no loss. If you want to go forward, you already have something built that can be utilized,” he said. “Even more, you can scale it up. A pilot is a powerful way for agencies to test and learn how AI technologies work.”
For Maximus, Beamon added, the company already has Federal Risk and Authorization Management Program (FedRAMP) certification for its cloud IVR and AI tools, so it can go from a pilot involving 3% of customer interactions to a program involving 100% of those contacts quickly.
Step 3: Expand use of real-time analytics
In many customer engagement centers today, a quality manager might listen to one or two phone calls a week, or even a month, involving a supervisor or quality manager to score and evaluate on how the agent performs. With AI and machine learning, it’s possible to score every call just like a human would, and, Beamon said, the results are nearly identical.
“They’re within a small percentage of each other from a scoring standpoint,” he said. “Why is this important? If I’m only scoring two calls for an agent that maybe does 100 or 200 calls in a week, am I really going to find where opportunities for better customer service could have been delivered or if they gave the wrong information?”
There’s also the ability to weed out fraud through analyzing those calls and other digital interactions in near-real time, Beamon said. “What AI/ML is doing is raising its hand because it’s listening to everything and saying, ‘Supervisor, I found something, come look at this call.’ Now, call center supervisors are looking at a much larger sample of data in real time that they can quickly act on.”
Plus, organizations now can evaluate trends and identify substantive ways to rethink customer services and experiences based on data and insights representing their specific agency’s constituencies.
“I’ve been in this business for 25 years, and the tools we have now are allowing us to be quick and responsive in ways that we never could before,” he said. “We’re going to see quality improve, and CX improve. We’re going to be able to deliver more services for the same budget. We’re going to see transformational things happening in the next five to 10 years.” (See sidebar, “A single digital front door to the government?”)
Step 4: Evaluate potential ROI
Beamon offered this take on what’s possible now: Applying robotic process automation is akin to having 1,000 people poring through the data. Likely, 95% of that data is probably benign. But perhaps 5% of it needs to be looked at.
“In the past, we were probably looking at 100th of the 5% that should have been looked at. Now, we’ve really shifted to where we’re looking at everything,” he said. “That means we’re focusing our human resources in on that 5%.”
Once an organization has a pilot in place, it can begin to look at how it can recoup dollars and put its resources into its biggest challenges, Beamon suggested.
“We have machines that can do things human beings can do for pennies per minute, and the humans are dollars per minute,” he said. “Right there is your value proposition of why you need to explore and investigate. For every minute that you move over to a digital assistant, you’re unlocking value in your operations. You are now taking that budget and putting it to work in other ways, which every federal agency can use additional budget to help them deliver more services.”
A single digital front door to the government? COVID-19 test kit program suggests the answer is ‘yes’
During the COVID-19 pandemic, Maximus took part in helping with the federal distribution of free rapid test kits to people nationwide. It was a project that allowed the company to implement interactive voice response, artificial intelligence and robotic process automation to integrate customer services at two federal agencies.
Although the Centers for Disease Control and Prevention spearheaded the free test kit program, the U.S. Postal Service fulfilled the orders.
“We actually built a solution where we had a contact center to support citizens so that they could order on the phone or go to a USPS website to place their orders,” said Maximus’ Andy Beamon. “A lot of people made phone calls. On each phone call coming to CDC, while we had a citizen on the phone, we were able to collect their information with a digital assistant and then create that order on the website using an application program interface.”
All of this happened instantaneously during a phone conversation between the caller and the digital intelligent virtual assistant to achieve complete self-service across two federal agencies. Even for callers seeking a live agent to assist them, the order would be placed in the USPS system and the call center agent would get a response in real time and be able to tell the citizen, “We’ve placed that order. You should be receiving it in a couple of days,” Beamon said.
Digitally connecting two agencies’ customer service experiences
Imagine the possibilities, he said. “There are agencies that work together, but people end up having to call one and then call the other,” Beamon said. “We have now achieved a next-generation use case for a citizen service involving multiple agencies. While we have people on a phone call, they never have to talk to a live assistant while obtaining cross-agency services. That is incredible.”
In the instance of the COVID-19 test kits, he recalled, sometimes people’s addresses did not match geographic location information maintained by USPS. The program allowed the CDC digital intelligent virtual assistant or live agent to open a trouble ticket with USPS, resolve the conflict with a citizen and ensure proper delivery of the test kits.
“We really are approaching an opportunity where one day in the future there might be one phone number that you call to have any problem you need resolved, or any service you want in the federal government achieved, through these solutions,” Beamon said. “We’ve done it now. We’ve connected two agencies to prove the art of the possible. What’s stopping us from connecting all the agencies and creating a single front door to the federal government?”