Chatbots and Trust: all you Need to Know

As a society, we trust chatbots to give us the answers that we expect. This is especially true for simple yet important interactions, such as confirming with an airline when a flight is, or adding a bag to the flight booking. But what about more complex or confidential situations? Booking a doctor’s appointment, applying for a credit card, or updating an insurance policy can involve divulging lots of personal information to a chatbot, such as your medical or financial history.
Can chatbots be trusted to keep information secure and offer good, relevant solutions?

The current technology landscape is becoming polarized by trust – with a collaborative economy driven by Airbnb and Uber, balanced by incidences of trust being betrayed. As several high profile examples have shown, at present, chatbots are not perfect. Despite the difficulties, there is a growing need for chatbots that can move beyond the simple interactions and can handle more complex and confidential information. Virtual assistants, chatbots that provide medical or financial advice, and even therapy: the potential for chatbots to improve customer experience and agent productivity is limitless.

One way to deliver this is through ‘supervising’ conversations. This is where every interaction by the chatbot has to be approved by an agent. After a training period, the percentage of interactions that have to be approved can decrease as the chatbot evolves to deliver better and better experiences. On top of this, agents can seamlessly pick up queries from chatbots if the need arises. This monitoring process boosts efficiency: agents simply have to approve a chatbot’s interactions, instead of spending time solving the enquiry themselves.