When the U.S. shut down in response to COVID-src9, many government agencies were overwhelmed by requests for help and information. One of those was Georgia’s Department of Labor (DOL) which had to deal with the unemployment rate doubling to 6.7% practically overnight. In business terms, this was a customer service crisis — and customer service is integral to customer experience.
“When things shut down and people were furloughed, they had to get unemployment and the phone lines were so crowded with people that they were eventually hanging on for hours and then being hung up on just because of the technical limitation,” said Donna Summer, operations and delivery manager for Digital Services Georgia (DSG).
DSG had a solution that wasn’t on a lot of peoples’ radar in 2020 but is now: Chatbots. Fortunately for all involved, the agency had been investigating chatbots for a while when this crisis hit.
“Even before COVID, we were researching chatbots,” Summer said. “Our motivation there is always to just meet our citizens wherever they are. We just want to make things easier for them and phone calls, emails take time. Some people that’s just not their mode. They prefer social media. Do they prefer a chatbot? So we were looking into it, but then, when COVID happened, we had to escalate our plans, really fast.”
Dig deeper: Chatbots could drive $srcsrc2 billion in retail sales by 2023
How fast? About three days. It was a very out-of-the-box implementation but it allowed them to automate the handling of a key customer pain point: Lost passwords.
“Can you imagine sitting on hold for three hours and not being able to log in and do whatever they had to do to get their unemployment?” she said. “So that was an immediate need and we’ve just kind of learned or are trying to learn how to optimize the chatbots from there.”
Now DSG is hoping other state agencies will see chatbots’ usefulness without being hit by a crisis. To that end, they’ve deployed one on their own site.
During their previous research, the agency put out a request for qualified vendors and become familiar with a number of companies. The one that stood out for them was Ivy.ai because you can limit where their chatbots get their data from. This limits the risk of “hallucinations,” when the AI invents answers.
“We build the bots using the customers’ sources of data,” said Sharon Harrison, Ivy’s vice president of marketing. “So it’s not bots going out and combing Google and wherever the hell else for information. You’re providing us with your sources of content — web pages, PDFs, spreadsheets, videos. And so the guard rails are that the bot is only looking for answers within that data set.”
For Summer and DSG, the use of chatbots is a customer service issue: How do you meet the customer where they are, when they are and in a way that’s easy for them to understand and use.
Dig deeper: How generative AI is improving customer experience and service calls
“Someone who works the night shift and they need information when nobody’s there, they can get it,” said Summer. “If they’re not someone who’s comfortable searching a website for information, they can just type in the question that they have, which is really nice. It’s on their terms.”
Harrison says that in addition to that, the chatbot is a force multiplier.
“Let’s say Kara’s on my marketing team and right now she’s surrounded by emails and phone calls,” said Harrison. “Well, if I can pull Kara away from that and let her actually use her skills to help with the implementation of the marketing strategy. I’d rather her be answering the questions that escalate and requires one-to-one attention versus the tier-one, low-hanging fruit stuff. That creates staff turnover because everybody burns out from those things.”
Hopefully, the increasing use of chatbots means organizations will already have the infrastructure in place to help their customers when the next crisis hits.
Get MarTech! Daily. Free. In your inbox.