AI-based chatbots are software that rely on technologies like natural language processing and machine learning.
By using smart contracts, integrating AI systems into customer support can have a significant impact and help mitigate potential issues.
AI in customer support enables faster interaction with users and improves the overall customer experience, but its systems require considerable time, large datasets, and a lot of work.
Chatbots analyse available data, identify patterns and relationships to make predictions, suggest decisions, or respond to user requests. However, their structures, logic, and data are often outdated. The case of Air Canada demonstrates the consequences of this.
An Air Canada chatbot suggested that the customer purchase a full-price ticket rather than a discounted one. It also informed the customer that they could receive the discount if attending a family funeral by filling out an online refund request “within 90 days of the ticket issuance date.”
Upon returning from the trip, the customer requested a refund, but an Air Canada employee replied that bereavement discounts were not applicable retroactively, citing terms from the website. The customer, who had saved the conversation with the chatbot, decided to sue Air Canada to recover the difference between the full and discounted fares.
Initially, the company, while admitting the mistake, attempted to deny responsibility for the information provided by its chatbot. However, the Civil Court of British Columbia ruled that Air Canada was responsible for all information provided, both on the website and through the chatbot.
According to recent research by Drift, the use of AI chatbots has increased by 92% from 2019 to today. Customers prefer AI chatbots for quick responses, but 34% feel frustrated when they can’t get an answer to a “simple question.”.
The imperfection of chatbots often stems from outdated information and data in the databases used to train them. If a query is phrased improperly or briefly, the chatbot may offer an illogical or incomplete response. Vague textual requests can lead to equally imprecise answers.
If Air Canada had used a contract-based logical structure, designed to respond to the regulatory obligations related to refunds, the chatbot’s autonomy could have been contained..
With the use of the Trakti platform for back office automation, the entire process becomes more transparent and faster.
Trenitalia has integrated AI into its customer support by adopting chatbots, blockchain, and Smart Refund, Trakti’s new end-to-end instant refund system, to monitor its customers’ travel status in real-time. The monitoring and certification of travel data, as well as the calculation of any refunds or vouchers to be issued, are automated. With Smart Refund by Trakti, the rules set by the contract and the law are automatically activated.
To discover how to digitalise your post-sale customer support process through AI, Blockchain, new Smart Contracts, and Open Innovation book a free demo.
