Most 'chat bots' still aren't smart enough to meet the high expectations usually associated with them, according to the results of a survey of 3,500 global consumers by Pegasystems. While most consumers agree that chatbots can be fast and convenient in certain situations, it is widely felt that they lack intelligence compared to human support agents, and 65% still prefer a human agent on the other end of a web chat. This article is copyright 2018 The Best Customer Guide.

Seventy-two percent of consumers generally find chatbots to be helpful to some degree, but the interaction quality can be quite mixed. The majority (58%) rank their chatbot experiences as merely 'adequate' - doing some tasks well and others poorly. Another 18% grumble how chatbots are ineffective or even annoying. Only 16% gave their chatbot experience a high quality rating.

Analysts expect chatbot usage to rise significantly in the next two years - a shift that could prove costly for brands that don't evolve their bots. According to Gartner, "25% of customer service and support operations will integrate virtual customer assistant (VCA) or chatbot technology across engagement channels by 2020, up from less than two percent in 2017."

"As chatbots become more pervasive, the quality of the engagement has lagged significantly behind customer expectations," said Ying Chen, head of product marketing, platform technologies, Pegasystems. "To truly depend on digital channels as the first line of defense in customer service, smart businesses need to unite their chatbots with the enterprise systems that can do real work - not just fetch bits of random information. At the same time, they must apply advanced artificial intelligence to deliver true personalized interactions in real time."

The best bots keep it simple
The results of the survey show that businesses still have a long way to go before consumers feel they can trust chatbots to give them an exceptional experience that can set a company apart. In fact, consumers tend to favour chatbots for only the simplest queries that can be done quickly. The most popular chatbot use cases are:

  1. Tracking an order (60% selected)
  2. Finding basic information (53%)
  3. Asking basic questions (49%)

The respondents also noted that fast service (56%), ability to engage on their own schedule (37%), and convenience (36%) are chatbots' top benefits. When done well, 43% say chatbots can be almost as good as interacting with a human, while 34% disagree, and 23% don't know.

Artificial intelligence or artificial stupidity?
However, speed and simplicity can only take today's digital consumers so far. While most brands claim artificial intelligence power their bots, consumers' top chatbot complaints include:

  • Not enough smarts to effectively answer questions (27%)
  • Lack of context in the conversation (24%)
  • Robot-like engagement with few human qualities (14%)

Similarly, the top reasons consumers would drop a chatbot session are when bots can't answer their questions (47% selected), make them do more work than expected (47%) or are too vague in how they can assist them (43%). Separately, only 17% said they would use a bot to purchase goods and services, further muddying the path from bots to direct revenue.

I'm in no mood for chit-chat
The survey found a separate group that have yet to try chatbots at all and have no plans to start anytime soon. These bot holdouts, which skew towards a slightly older demographic, say they haven't used chatbots yet due to:

  • Lack of any real exposure to chatbots (53%)
  • A personal preference to only engage with a human (30%)
  • Lack of knowledge on how to use chatbots (23%)

A full 45% of these consumers without chatbot experience said they won't try one in the next year while another 30% aren't sure, which presents a roadblock to companies trying to expand reliance on digital service channels. Only 25% of these non-users said they would be willing to experiment with a chatbot, albeit with some reluctance. Their biggest concerns stem from their lack of experience: they simply don't know how to use chatbots (top concern at 46%), lack confidence in chatbot effectiveness (31%), or worry about security and privacy (27%).