Research undertaken by BT seems to suggest so.
With 82 per cent of consumers saying it would be good if organisations used AI to monitor the condition of products and services; and 72 per cent wanting AI to listen in to calls and intervene where help was required.
Considering the highly publicised issues in recent years to do with data privacy, this is quite surprising. Whichever side of the debate you advocate — more listening by machines or less — if your organisation is considering using machine learning and AI to be more proactive in how you reach out to customers, there are some fundamental questions you need to ask before making the investment case.
According to MIT's Sloan , investment decisions related to AI-type digital technologies need to be predicated on whether the technology enables the organisation to: “interact with customers in more meaningful ways by providing ever-increasing personalised, convenient, hassle-free transactions”; as well as whether digitally enhanced solutions “…help solve customer problems”. In this context, a digital enhancement means equipping products with real-time data and making them smart (e.g. installing software, operating system, sensors and wireless connectivity into a washing machine).
Where we see companies implementing AI to “interact with customers in a more meaningful way” and “help solve customer problems” is often in situations where there is a predefined workflow for the machine learning algorithm or AI to follow.
For example, one airline customer we work with wanted to do the following. When some of their passengers were faced with travel disruption due to bad weather, their contact centre was flooded with incremental calls which they couldn’t resource. Instead, they wanted to send an outbound proactive text-based notification to their passengers’ mobiles. Text was more likely to get through than using an internet-dependent tool such as WhatsApp. The SMS notified the passengers that the flight was cancelled due to bad weather, whilst informing them they had been booked on the best alternative flight, with its details. If the passenger agreed, they would simply reply with a ‘1’ and would then be sent their booking details and boarding card. If they didn’t agree, then ‘2’ would initiate a chatbot conversation running an interactive text response session with the passenger. The chatbot followed a predefined workflow to come up with an alternative option for the stranded customer. Should the passenger still be unsatisfied, then the real-time customer journey analytics tools running in the background would route the query to a human agent to remediate the issue.
Elsewhere, a debt recovery company used real-time speech analytics for indicating in a live call, when vital or mandatory information had been missed out by the agent. The agent was then prompted about this. This feature helped them meet their compliance requirements and avoided misunderstandings with customers in the future. By listening in, the AI running the real-time speech analytics monitored the customer’s and the agent’s speech, providing live feedback to agents as well as supervisors. Where an agent needed support, a supervisor was quickly marshalled to provide them with reinforcements or, if needed, intervened in the actual engagement with the customer.
For every global corporation the value of their brand is too important to risk it being affected by a poor service experience, especially if it’s the result of a cost-cutting exercise to remove human agents. Where there is limited potential to get things terribly wrong in a service experience, consumers welcome the opportunity for automation, but in an important or sensitive situation and as a fallback there always needs to be a live human agent.
 J. W. Ross, I. Sebastian, and N. O. Fonstad “Define Your Digital Strategy – Now,” MIT Sloan Research Briefing, 6 June 2015, Volume XV, Number 6