“I’m Sorry, I Didn’t Get That”: Flipping CX Perspectives with Conversational AI
September 14, 2022
A brief history of AI
The concept of conversational artificial intelligence (AI) has been around for over 50 years, reaching back to the first primitive chatbot, ELIZA, which was developed from 1964-1966 at MIT by Joseph Weizenbaum. Similar to modern chatbots, ELIZA was designed to query and respond to user inputs based on programmed scripts, though back then there was no such thing as machine learning or natural language understanding or processing. ELIZA later inspired Richard Wallace to develop A.L.I.C.E. (Artificial Linguistic Internet Computer Entity) in 1995, which was the first attempt at Natural Language Processing (NLP) technology, and the invention of AIML (Artificial Intelligence Markup Language, also AI/ML), which is still used today..
As you might imagine, technological advancements in NLP have skyrocketed since then, bringing AI interfaces such as Alexa, Siri, and Google Assistant into our homes, resulting in chatbot technology that allows us to query knowledge bases and manage services with the brands we purchase from and interact with every day.
Perception and Adoption
Despite exponential advancements over the last decades, consumer trust of this technology in the customer service (CX) universe has been understandably slow, with 86% of surveyed respondents preferring human interactions over chatbots, as reported by Forbes in a 2019 article. An interesting detail of those survey results is that the number of people who would choose chat for quick answers over other channels had actually reduced from 50% in 2018, to only 29% in 2019. One could postulate that 21% didn’t appreciate their experience in the interim period between surveys.
What is profoundly interesting is the impact the global pandemic had on both consumer perceptions of AI and brand adoption across customer service channels, and how this contrasts with pre-pandemic sentiments. Fast-forward to 2022, this Zendesk survey indicates 69% of consumers would choose a chatbot for simple issues, a change of 40% (of the total surveyed) in three years, as well as a 23% increase over the previous year. On the flipside, 37% of IT professionals stated that AI was already critical to business in 2019, and this number has only increased, with 78% of professionals in top companies stating that AI is a key revenue driver in 2022. Now, perceptions and acknowledgements are one thing, but adoption of AI technologies across customer service channels has yet to catch up, with only ¼ of those same companies having actually scaled their AI initiatives.
Setting the Human-Machine Balance
Consumers have been following CX trends towards automation with some trepidation over possibilities that machines will one day entirely replace the human element. Don’t worry, that’s not likely to happen any time soon as it seems 71% of consumers would avoid brands that don’t have human agents, plus there will always be situations that only a human touch can resolve.
Regardless of whether service is provided by artificial intelligence or a human resource, one thing is apparent, and that’s the need for interactions to have a human feel to them. Nobody really desires communicating with a robot, so that’s where conversational AI can take centre stage.
While many companies have been quick to adopt AI technologies for self-service options, there’s still work to do in many cases, as seen in this BMW forum from 2021 where customers decry a new IVR system’s ability to handle even the most basic requests regarding the status of their vehicle shipments. Between reported poor audio quality, the IVR’s inability to comprehend basic requests and statements, plus lengthy wait times—often resulting in call terminations, customers were left justifiably frustrated by the experience.
It’s easy to swiftly settle on the latest technology solutions, but it may not be the wisest to implement them without first deciding how those important customer-company interactions should look. Companies with contact centres, web-based touchpoints like contact forms and social media properties, as well as email feedback, already have a mountain of data they can scour and analyze to understand not only what customers want from their relationship, but also how they communicate during those interactions. This is key information that can help immensely with the design, testing and implementation of conversational AI touchpoints to provide a more robust and streamlined experience. The key benefit of AI is not simply mastering a single method of contact, but to apply holistically, so that no matter how a consumer interacts with your business, they are getting exactly what they need, in a way that is more empathic and personalized for their specific needs and experience.
If we return to the example of the newly-introduced IVR system that BMW implemented, we can see that while the intent was to provide a more personalized service to customers checking on their orders, there was a missed opportunity. How might the outcomes of those interactions have changed with a more careful approach to AI design? One comment in the forum highlighted that the system’s voice had a new British Accent, which lends the question, was the system designed to process the voices and accents prominent across the USA, or was it designed for British accents? If they had accounted for this, plus all the possible responses or requests their customers would speak, this could have resulted in more adequate query handling, less frustration for customers, while completely removing wait times for unnecessary agent transfers. Not only would this have improved customer perceptions of their CX system, but it would have reduced the load on their contact center systems and dropped fewer calls—reducing pressure on their bottom line, and thereby increasing the profitability of their CX operations.
Additionally, the implementation and promotion of a chatbot knowledge base could have saved customers from having to call into the system in the first place. A simple query for status using their order confirmation numbers would have immediately returned the information they desired, using only a mobile phone, or a desktop computer.
Maintaining a fluid strategy that adjusts for both changing customer needs and technology advancements is increasingly important these days. While it’s generally easy to gain new customers, it’s not so easy to keep them if there’s a mismatch between customer needs and a company’s perception of them during their customer experience lifecycle.
Understandably, not all organizations have the budget scope to immediately implement omnichannel solutions, but with strategic planning, and a good understanding of where and how their customers prefer to interact with their brand, even small companies can make a big impact with CX.
Are you ready to make an impact with your customers, using conversational AI? Find out how Connex can help your organization by sending us a message, today.