Digital workers are invading the enterprise -- and that's not a bad thing
Credit Suisse and LogMeIn executives provide a glimpse at how external and internal customer service strategies are melding.
CIOs who want to prove out the value of artificial intelligence in the enterprise and modernize their ITSM portfolios at the same time should take a page from the playbooks of two global companies.
In an effort to serve internal customers more efficiently, the head of cognitive and digital services at Credit Suisse Group AG, as well as the CIO at LogMeIn Inc., have introduced digital workers to the organization. These digital agents operate as a first point of contact on the service desk, taking on simple, repetitive tasks such as resetting a pin or providing basic regulatory information for employees.
The invasion of digital workers in the enterprise is part of a larger trend that not only reflects how cutting-edge technologies like machine learning and natural language processing are breathing life into stodgy processes, but also how IT executives are cribbing from their company's customer experience strategy.
An investment in efficiency
Jennifer Hewit began managing the global services desk at the Zurich-based Credit Suisse in 2016, and she said it was clear from the start that she'd have to transform how it operated. "The only way [employees] were able to access support was via voice," said Hewit, head of cognitive and digital services at the financial institution. And by voice she meant the telephone.
That meant if, say, a major snowstorm prevented employees from getting to the office, they'd have to call the service desk and talk to an agent who could spend as much as 40 minutes trying to help them gain remote access to applications and systems.
Hewit was researching products that could help her introduce chat to the financial institution when she came across IPsoft Amelia. Branded as "the most human AI," IPsoft doesn't refer to Amelia as an assistant or a bot, but instead calls it a digital agent or colleague that can understand and respond to what's being said and how it's being said.
Hewit liked that IPsoft's digital worker was a step beyond a live chat channel: Amelia could take on lightweight, repetitive tasks, freeing up human agents to concentrate on more complex issues. She and her team, which included a neuroscientist and a computational linguist, started by "building Amelia's brain." They began to breakdown the top 10 call drivers that agents typically solved within the first 10 minutes of a call in an effort to automate them.
And the efforts have proved fruitful, as demonstrated by Amelia's ability to respond to an outage that locked 4,000 employees out of a critical application last February. The digital worker was able to reset 67% of the pins, giving live agents a chance to focus on more technical problems, according to Hewit. "For me, Amelia was a saving grace in that scenario because she was able to work alongside her human colleagues to offer a better user experience."
Amelia serves as the first point of contact for employees. If the digital worker can't provide assistance, she will pass the request along to a live agent. Chats between Amelia and employees and live agents are recorded, and the data is used to refine and bolster Amelia's capabilities, according to Hewit.
A regulatory robot
The IT bot at LogMeIn, a software-as-a service company in Boston, is called Bishop, named after the friendly android in the Aliens series. Bishop can be accessed by employees through the collaboration tool Slack, and it provides information on products, internal processes, as well as information about security and regulations -- topics CIO Ian Pitt is hoping Bishop can master.
"As you can imagine with the world being given GDPR this year, there are a lot of questions from sales and legal and customer care -- because everyone is asking the same questions," Pitt said. "So the next rollout is to take what we've already got and expand it to cover more security regulations."
Getting "a critical mass" to use any new technology can be a challenge, and Bishop was no different, according to Pitt. In the early days, the novelty of Bishop was enough to attract employee attention. "After the initial excitement died down, we saw some drop off," he said. "So we then started working hard on increasing the content that goes into the chatbot and making sure it was maintained."
For example, Pitt and his team focused on the typical queries that arise from a merger and acquisition. Employees often have basic questions such as where to get laptops from or what kind of applications they can use, and they can turn to Bishop for answers.
Pitt said Bishop's main metric is what he called "case deflection," any instance where an employee gets a complete answer from the digital worker rather than having to raise it to IT or take it further. "We're currently running at about 85% to 90% accuracy," he said. "That means the help desk team can go work on the real problems that only a person can fix at this point."
The disappearing distinction
The technology behind LogMeIn's Bishop came out of an acquisition of Nanorep, an Israeli chatbot company. And it's also used to power LogMeIn's Bold360, a customer engagement product.
"When we got Nanorep, we wanted to roll it out to, one, deliver the service to customers, and, two, to do what we called dogfooding, which is essentially using our own products to improve our own users' experience," Pitt said.
More than dogfooding, Pitt provides an example of how the distinction between a strategy for internal versus external customers is disappearing. "More and more as we're focusing on employee engagement and the importance of employee engagement, they're really two sides of the same coin," said Kate Leggett, vice president and principal analyst at Forrester Research.
Leggett said customers, regardless of whether they're internal or external, want easy access to information, an easy resolution to their problem, and self-service as their first point of contact.
Digital workers such as Amelia and Bishop take things a step further: They are examples of self-service channels powered by AI and automation, which Leggett said is a big customer service trend. They're filtering out simple, streamlined inquiries and escalating more complex tasks to live agents.
"Which means that your human agents may have to be upleveled, they have to be coached differently, they may have to be educated differently," she said. "The nature of work is really changing."