A study of over 2,500 consumers CEOs and other senior decision makers on artificial intelligence (AI) spending at some of the largest communications and media companies in the region revealed how plans for a rise of the robots for frontline customer interactions could be hampered by wrong investment choices and a lack of human talent. These are part of the findings from a commissioned study conducted by Forrester Consulting on behalf of Amdocs in the Americas, Europe and Asia.
Nearly a third (31%) of consumers in Asia interact with virtual agents at least once a week. While 46% say it is convenient and saves time (46%), 47% also say they had no other option. If offered a choice, 83% would prefer to speak to a human since human agents better understand their needs (81%) and can address multiple questions at once (60%). Bots, say consumers, cannot deal with complex requests (their biggest problem), deliver personalised offers as well as humans (the second-biggest problem) or understand human emotions (the third-biggest problem).
Consumers also have strong views on how they want bots to look like and behave. Forty-eight percent prefer their bot to look like a human, as opposed to 20% who want to see an avatar. While 43% of consumers are neutral, 42% prefer them to be female, rather than male (15%). In order of importance, sounding polite, intelligent and caring were ranked the highest in terms of preferred bot personality traits, followed by being fully customisable and funny. Conversely, sounding serious and authoritative ranked much lower than other attributes, receiving 14% and 12% of the vote respectively.
Nearly a third (31%) of consumers in Asia interact with virtual agents at least once a week. While 46% say it is convenient and saves time (46%), 47% also say they had no other option. If offered a choice, 83% would prefer to speak to a human since human agents better understand their needs (81%) and can address multiple questions at once (60%). Bots, say consumers, cannot deal with complex requests (their biggest problem), deliver personalised offers as well as humans (the second-biggest problem) or understand human emotions (the third-biggest problem).
Consumers also have strong views on how they want bots to look like and behave. Forty-eight percent prefer their bot to look like a human, as opposed to 20% who want to see an avatar. While 43% of consumers are neutral, 42% prefer them to be female, rather than male (15%). In order of importance, sounding polite, intelligent and caring were ranked the highest in terms of preferred bot personality traits, followed by being fully customisable and funny. Conversely, sounding serious and authoritative ranked much lower than other attributes, receiving 14% and 12% of the vote respectively.
The research has shown that service providers priorities are not investing in the right areas in terms of their AI investments. Thirty-eight percent of service providers in Asia are prioritising AI investment in increasing information security and privacy and 33% are focusing on speed of response. Yet customers would rather have bots deliver better personalisation and more comprehensive information. Both are lower on service providers’ priority lists at 8% and 4% of respondents saying they are important.
Thirty-one percent of service providers are also creating avatar images for their bots whereas consumers prefer human-like images. They are also investing in undesirable features - more than half (56%) of service providers are building bots that sound serious and nearly a quarter (24%) are making them sound authoritative. The fact that 39% of consumers experience problems with today’s bot services but do not complain about it means the industry is working in the dark, Amdocs said.
Eighty-seven percent of service provider AI decision makers in Asia say that 85% of customer interactions will be with software robots in five years’ time. And almost a half of these decision makers (46%) fear they are lagging behind their competitors in the use of AI to improve the customer experience. To catch up, over a half (58%) plan to increase their AI budgets by at least 6% in the next 12 months and 87% intend to expand their AI workforce within the year.
But this expansion might still not be enough, Amdocs said. Contrary to the common perception that tomorrow’s service provider will be run more by robots than humans, only 4% of service providers see AI as the opportunity to replace a large number of staff. Most decision makers (67%) actually see the lack of human skills to set up and run AI as one of the top two risks to delivering on their AI strategies, just behind technology not being mature enough (79%). A third are seeking external support, predominantly from their existing vendors as opposed to native AI solution providers (25% vs 8%).
“Consumers have a good sense of how bots can serve them, better-developed than perhaps the industry’s. Their level of frustration with today’s bots is striking; a third even say they will take their business elsewhere if the poor service continues,” says Gary Miles, GM, Amdocs. "The good news is consumers actually believe that if anyone can get AI right, the communications and media industry can. And that’s ahead of banks and retailers. So AI could be a winning gambit for service providers as long as they sync up their AI investment priorities with what customers actually want.”
“The research shows, however, that many service providers do not believe they will be able to achieve this on their own,” added Miles. “They are mostly turning to their existing vendors and not to native AI solution providers, probably in order to ensure AI does not become another tech silo that is hard to scale and manage."
The research covered consumers and senior service provider decision makers in Asia, providing a wealth of market-, gender-, age- and role-based information. An equal mix of 2,575 female and male consumers between the ages of 18 to 74 were surveyed, as well as ten service provider executives from the region’s top communications and media service providers, half holding C-level roles.
Interested?
