“Bots are certainly the first thing people think of, in terms of AI in customer service”, admitted Steve Nattress, Product Director at Enghouse Interactive. “But while helping customers with self-service has a big impact, there’s a whole other side to it, that’s really growing in importance.”
And although self-service bots are increasingly sophisticated and responsive, we’re generally aware — when we stop to think about it — that we’re interacting with an AI, rather than a human. Provided it meets our needs in the moment we really don’t care, and the majority of our sales and support needs are predictable and straightforward as a customer journey. It’s those edge cases where we need the human touch, and miss it if it’s not there.
The human touch of empathy
“The key part of good customer service is empathy. Building rapport with the person you’re talking to, building trust. That part of the human relationship, AI can’t do — yet,” Nattress continued.
But AI can help us by recognising and identifying empathy, or recognising its absence — flagging that for attention, either in real-time from a supervisor, or for future remediation, and improved outcomes — thanks to natural language processing.
“During the conversation, it recognises intonation and content. And post conversation it can analyse the transcript, looking for additional insight beyond just simple keywords and phrases. The AI can spot where things went wrong, in ways that weren’t possible just a couple of years ago, including detecting subtle shifts in emotional responses — similar to the way humans do so instinctively, but at an unprecedented scale.”
A silent assistant, in the moment
The AI supervisor can offer practical help too, during calls or live-chat — flagging questions or commitments yet to be followed up, or retrieving required answers and documentation responding to conversation triggers from either party. The best advice and answers can be provided at record speed, reducing the time the enquirer spends on the interaction and increasing first call resolution and customer satisfaction metrics. “It’s really cost-effective to implement this functionality because the return on investment is about improving the efficiency of the agent.” Nattress explained. “Making sure the agent is efficient, and also consistent and accurate — these things really drive the fundamentals of good customer experience.”
This definition of CX explains why chatbots have sometimes led to dissatisfaction, if they are used in inappropriate circumstances, or deployed with inadequate supervision and ongoing development and review — which is no longer excusable when the machine learning functionality is there to continually improve their scope, as well as to identify and promptly route out the interactions which need it.
And this learning is supported by another thing that AI can do better than any human, which is, crunch the data — generating predictive insight from vast datasets and helping contact centres improve their metrics continually, through zeroing in on the right tweaks.
Big data, small shifts
“Our AI solution is different to others because of natural language processing,” Nattress described, explaining how the combination of accurately inferring emotions combines with the capacity to analyse and synthesise conclusions in a way no human brain could conceive.
“You can do cross-analysis on hundreds of thousands of recordings, and identify the exact moments of the interaction to change, where phrasing a question slightly differently could perhaps produce a dramatically different response in sales volume or NPS”
So, the future of AI in the contact centre isn’t about agents being replaced by chatbots all together, it’s about the symbiosis of human empathy with AI-driven data analysis, to continually optimise the experience. And not only is this a win-win for enterprises and their augmented CX agents, but it’s also excellent news for the most important people in the whole interaction — the customers.
