We’ve all made that angry call to customer service in a moment of panic or sheer frustration only to be transferred to the wrong department or be forced to repeat our grievances to multiple customer service agents. Whilst we all want to talk to a ‘real person’, conversational AI tools are getting much more sophisticated and can provide real benefits to customer interactions.
From a business perspective, voice AI more specifically is expected to reduce global call centre labour costs by as much as US$80 billion by 2026, according to Gartner.
These days, conversational AI tools can capture each customer's conversation and accurately document it for enhancing future interactions. It can even use speech analytics technology to analyse sentiment during calls so that customer service agents can be brought in when a customer sounds distressed or angry. Conversely, it can detect gratitude or joy to help determine an organisation’s overall customer satisfaction levels. It can also help protect sensitive customer data to meet compliance needs by detecting and redacting information in real-time.
Despite the increased capabilities and benefits to both an organisation and its customers that conversational AI now provides, in Australia, the adoption of conversational AI for customer service remains relatively low compared to other countries. In saying that, APAC is expected to see the highest growth by 2024. Today local organisations remain more interested in web-based chatbots and automated interactive voice response(IVRs), which remain the primary AI systems used in Australia.
So, what’s holding conversational AI back?
According to Gartner, roughly ten per cent of agent interactions will be performed by conversational AI by 2026. This is compared to only 1.6 percent of interactions being automated in 2022.
The first reason is that legacy infrastructure doesn't necessarily support real-time or call recording in stereo which is necessary to differentiate between caller and agent in conversations. It also doesn’t allow for real-time processing or recording of calls across all voicetouch points, including contact centre technologies and unified comms platforms like MS Teams and other apps. For example, when an agent answers a call in thec ontact centre, they may have an on-premise solution, but what happens when the call needs to be transferred to the relevant department via MS teams? Only part of the conversation would be recorded and processed. This is where SecureCo’s Intelligent Voice Platform comes in. By orchestrating voice calls across all customer touch points via secure channels, the platform can record and process information in real-time, meaning you can ensure your AI investments are being leveraged across all platforms.
There is also a misconception that organisations need to spend a lot of time implementing the technology and training AI for specific use cases, but that’s largely not a concern anymore. This is because the data sets are getting larger and more industry-specific libraries are becoming available by companies that have already done the leg work.
Lastly, previous bad experiences with not-so-sophisticated AI technologies have left a bad taste in many people’s mouths. It’s because of these experiences that many say they prefer to speak to a ‘real human’, with a study conducted by PwC finding 59% of consumers feel companies have lost touch with the human element of CX.
To make things worse, the pandemic took away our ability to have face-to-face interactions with brands. This meant consumers were increasingly looking for a human connection, and with many finding themselves in difficult or uncertain circumstances, needed the empathy of a customer service agent.
Mixing AI with empathy
Given some of the hesitations around the uptake of conversational AI, an all-or-nothing approach doesn’t make sense. Organisations looking to create efficiencies and cost savings with AI are going to need to temper the implementation of the technology with human empathy.
Organisations need to create a plan around what parts of each customer experience touchpoint will be automated and which will be handled by agents. This could be based on the complexity of the enquiry or the type of customer, for example.
It’s also important to base these decisions on research. Before implementing AI, organisations need to have a firm understanding of their interactions with their customers. Listening to calls and capturing caller intent will support in gaining the insight they need to decide what to automate and when and the amount of resources required at each touchpoint.
Past decisions to shift contact centres overseas and rely on unsophisticated AI technologies have meant Australian organisations have some catching up to do when it comes to providing the customer experience that consumers now demand. Today’s conversational AI technology has the ability to meet these demands, but organisations need to be thoughtful in how and when they implement it so that the transition is seamless and the customer sees the benefits. By ‘augmenting humans’ rather than replacing them, conversational AIcan be a game-changer for contact centres, saving on costs whilst providing an enhanced customer experience.