Julien Rio, a Certified Customer Experience Professional (CCXP) and AVP of International Marketing at RingCentral, is a renowned figure in the CX world. Julien is the founder of CX Therapy, a vlog that delves into real-world customer experiences, and the author of “Customer Experience Unearthed.” Recognized for his contributions, including being named among France’s top marketing influencers, Julien’s passion for storytelling and customer engagement marks him as a respected leader in the field. As a founding member of the European Customer Experience Organization (ECXO), he continues to influence and inspire with a commitment to advancing customer experience.
This week, the buzz surrounding Air Canada’s chatbot reverberated across discussions, with many scrutinizing its overzealous behavior in offering non-existent discounts to Jake Moffatt, a passenger who promptly booked a next-day flight based on this erroneous information.
In concrete terms, the incident traces back to 2022 when Jake Moffatt, grieving the loss of his grandmother, turned to Air Canada’s website to secure a ticket for the impending funeral. The chatbot assigned to Jake’s inquiry provided misleading guidance, erroneously outlining the discount claim process. Subsequently, when Jake attempted to avail the discount post-funeral, Air Canada rejected his request, citing failure to adhere to the correct procedure.
In response, Jake pursued legal action, prompting Air Canada to assert that the chatbot operated as a distinct legal entity, responsible for its actions. However, on February 14, 2024, the court ruled against Air Canada, compelling the company to reimburse the discount, along with covering legal fees and damages.
What implications does this have for chatbots?
This legal precedent underscores a fundamental truth that resonates with consumers: chatbots are integral components of a company’s infrastructure, akin to websites, FAQs, advertisements, or staff communications. The expectation is clear: consumers should be able to place trust in the information provided by chatbots without feeling compelled to verify every assertion.
But does this ruling signify that chatbots pose inherent dangers? Should companies reconsider their reliance on them?
The landscape of chatbot technology is evolving rapidly, particularly with the proliferation of AI and Generative Pre-trained Transformers (GPTs), marking a departure from traditional decision tree-based models. While these older models boasted simplicity and accuracy, they often hindered direct communication with human agents and could occasionally mislead consumers and generate frustration. The emergence of AI ushers in a new era for chatbots, enabling them to autonomously process vast amounts of data, analyze inquiries, and provide nuanced responses in real-time. However, this newfound autonomy comes with risks. As evidenced by numerous GPTs, there is a propensity to fill knowledge gaps with fabricated information and a lack of discernment in critical scenarios.
So, what does the future hold for chatbots?
Should companies abandon chatbots and AI altogether?
Absolutely not!
Similarly to human customer service representatives or store employees, chatbots and AI are prone to errors. However, there are two significant distinctions. Firstly, human staff members are, well, human – it’s expected that they will make mistakes. It’s unrealistic to anticipate that humans can sift through billions of datasets to find answers or retain every conceivable response. Occasional errors are part of the human condition, and companies typically do not attribute individual actions to being “a separate legal entity.” For instance, consider the “Norwegian Sovereign Fund,” which recently reported an $86 million loss due to a human error in an Excel sheet. In response, the CEO acknowledged the mistake with the understanding that such mishaps are inevitable in human endeavors.
Secondly, human employees are guided by company policies and protocols. They understand the rules, the potential consequences of their actions, and strive to prevent errors that could adversely impact the organization. In contrast, chatbots and AI lack this consciousness. Their primary function is to provide rapid and accurate responses, devoid of human-like consideration or understanding.
While chatbots and AI may not possess the same level of consciousness or adherence to guidelines as human employees, they still offer immense value when properly utilized. Rather than viewing them as replacements for human interaction, companies should consider them as complementary tools to enhance efficiency and customer service. By acknowledging their limitations and implementing appropriate safeguards, companies can leverage chatbots and AI to streamline processes and improve overall performance.
What’s the optimal approach, then?
Much like you wouldn’t hire a new employee and leave them unattended in a store or call center to field customer inquiries without proper training and oversight, chatbots—especially AI-based ones—require meticulous training, control, and maintenance. So, where do we begin? Let’s break it down into four key steps.
Defining the Scope:
Start by pinpointing a specific scenario. While AI-based chatbots theoretically have endless capabilities, it’s prudent to start with a concrete use case. For instance, through customer journey mapping and recent surveys, identify recurring issues like billing inquiries. Training the bot with a vast array of examples—drawn from hundreds or thousands of customer service interactions—provides a solid foundation. Educate the bot on three crucial principles: only respond to questions it’s trained for, transparently admit when it lacks an answer, and facilitate seamless handoffs to human agents when necessary.
Opening in Waves:
Once initial training is complete, gradually introduce the chatbot to a limited audience. Begin internally, testing its responses to both simple and challenging queries, and making adjustments as needed. Then, gradually expose the bot to a select group of customers, closely monitoring its performance. Be proactive in addressing any mistakes, intervening directly with customers if necessary. As confidence in the bot’s capabilities grows, gradually expand its reach. However, maintain control by setting clear expectations and ensuring users understand the bot’s limitations.
Monitoring:
As the bot interacts with a broader audience, ongoing monitoring is essential. Similar to how contact centers have supervisors to support agents, chatbots cannot operate autonomously. Continuously review conversations, intervene when issues arise, and refine the bot’s instructions to enhance its performance. Rigorous monitoring is crucial for maintaining quality and addressing any emerging challenges.
Expanding:
Once the bot demonstrates satisfactory performance within its defined scope, solicit feedback from customers to gauge satisfaction. If deemed successful, consider expanding its capabilities to address additional scenarios. Utilize the same methodical approach to identify new areas for improvement and begin the training process anew.
By following these steps, companies can methodically deploy and refine chatbots to enhance customer experiences and streamline operations. With careful planning, training, and ongoing monitoring, chatbots can serve as invaluable assets in delivering efficient and effective customer service.
Is AI all about Chatbots?
AI encompasses more than just chatbots. Artificial Intelligence presents a vast array of possibilities beyond chatbot functionality. Consider this: wouldn’t it be remarkable for your customer service representatives to have an AI-powered assistant whispering the right answers in their ear? Agent assist technology, fueled by AI, empowers your agents with superhuman capabilities while mitigating the risk associated with AI being at the forefront of customer interactions. Furthermore, AI can revolutionize customer inquiry management by accurately understanding and routing messages to the most qualified and available agents. It can also aid in predictive analytics, preemptively resolving issues before customers even become aware of them, addressing a common complaint—lack of proactive communication.
Indeed, Artificial Intelligence offers myriad ways to enhance the customer experience. While chatbots represent a visible manifestation of AI, they are merely the tip of the iceberg.
The tale of Jake Moffatt and Air Canada may instill fear in businesses considering the deployment of chatbots and AI. However, rather than succumbing to apprehension, it’s vital to glean valuable insights from such incidents. As Artificial Intelligence becomes increasingly prevalent, stories like these may proliferate.
Yet, there’s no need for panic or retrenchment. Instead, a customer-centric approach is imperative. Viewing chatbots as integral members of the workforce—requiring onboarding, training, and diligent oversight—ensures alignment with customer interests. When deployed thoughtfully, AI-based chatbots have the potential to revolutionize the relationship between your company and its customers, serving as a formidable catalyst for positive change.