Chatbots have the potential to reduce the load on contact centres. A recent Capita workshop explored the impact of this and other emerging technology.
Earlier this month, Transport for London (TfL) unveiled a chatbot integrated with Facebook Messenger to reduce the volume of queries going through its contact centre and offer interactive help for the travelling public.
TfL’s so-called TravelBot can “chat” with customers using Messenger and instantly tell them when their bus is due to arrive, provide service updates and Tube maps.
The TravelBot is linked directly to a customer service agent for more complex queries, and has the ability to “learn” the correct way to answer queries.
According to analyst Gartner, specialised chatbots could reduce the number of helpdesk workers by enabling routine requests to be handled via automated chatbot responses. “Some responses can be automated via scripted replies, while others will merely do a better job of placing callers on a decision tree and will ultimately be handled by human respondents,” Van Baker and Magnus Revang wrote in the Gartner report Four use cases for chatbots in the enterprise.
Chatbots are part of a wider trend that is driving the adoption of artificial intelligence (AI).
A Deloitte University Press article by William Eggers, David Schatsky and Dr Peter Viechnicki, published in April 2017, predicted that AI would spawn huge changes in the public sector, transforming how government employees get work done. The authors expected AI to eliminate some jobs and lead to the redesign of countless others, as well as create entirely new professions.
“In the near term, our analysis suggests, large government job losses are unlikely,” they wrote. “But cognitive technologies will change the nature of many jobs – both what gets done and how workers go about doing it – freeing up to one-quarter of many workers’ time to focus on other activities.”
In the US, the Department of Homeland Security’s Citizenship and Immigration Services has created a virtual assistant, Emma, that can respond accurately to human language. According to Schatsky and Viechnicki, Emma uses its intelligence simply, showing relevant answers to almost 500,000 questions a month at present.
“Learning from her own experiences, the virtual assistant gets smarter as she answers more questions,” they said. “Customer feedback tells Emma which answers helped, honing her grasp of the data in a process called ‘supervised learning’.”
Such applications can be more intuitive compared to a frequently-asked-questions page on a government website, but the real power is when the chatbot conversation is linked to back-end systems. “In this situation, a change in a back-end record will trigger an event, which can cause a message to be delivered to an enterprise messaging or workflow environment requesting an employee response,” wrote Gartner’s Baker and Revang.
Public sector uses
This week, Computer Weekly attended an interactive workshop run by IT services firm Capita looking at the use of chatbots and headless interfaces in the public sector.
Among the demonstrations was a chatbot used to fire off the business process resulting from a missed bin collection. This is one of the areas where Capita has been looking at how people can interact with their local authority electronically using voice-based queries. It is called a headless user interface.
Stewart Davison, head of business development at Capita, said: “Last year, Siri and Cortana were probably fringe technologies, but Amazon Echo is probably more mainstream.”
He said councils had spent a lot of time and money on websites, but few people use them. “A large proportion of the public is not digitally engaged,” he said.
A headless UI such as the Amazon Echo Dot offers local government a possible way to overcome this barrier. One of the examples Capita showed during the workshop was voice access to a social housing portal. Davison said: “One of the reasons people ring up the council is for repairs. There are hundreds of calls a day and logging repairs requires manual intervention, even though there are back-end systems.”
Local authorities could link an Echo Dot with home automation such as internet-connected smart switches to support vulnerable people. One council in Scotland is considering using the Echo Dot to provide a low-cost alternative to personal alarms used to support elderly people if they have an accident at home.
The Capita workshop showed that although the public sector sees the benefits of chatbots and headless user interfaces such as Alexa, there is need for collaboration.
Davison said: “It has only been a few months since Alexa has been out. It is an emerging technology and so there won’t be case studies yet.”
This means that to develop a strategy around these technologies, it is necessary to collaborate across healthcare, social housing, academia and the charity sector, among others, as well as with consultants and even a totally different sector, said Davison.
Among the examples discussed at the Capita workshop was how Alexa could be deployed for internal processes, such as booking meeting rooms or checking the availability of a hot desk.
Another example, from the charity sector, looked at the possibility of using chatbots to extend the service the charity offers from a nine-to-five, one-on-one telephone service that handles 7,000 calls a year, to a 24/7 service handling up to a million enquiries.
Chatbots and headless user interfaces could help the public sector to streamline digital services. But with public sector services underfunded and coming under increasing pressure, adopting these technologies requires a strong return on investment. “You need to look at your own processes and use these as a baseline on which you can attach a value,” said Davison.
For instance, the cost justification of adding a chatbot can be calculated by assessing the cost of handling common queries, he added. “What are the top five simple questions being asked in our call centre? You could attach a cost per telephone interaction, and compare this to the web channel.”
As published by Computer Weekly – view the article here