AI: the human touch in customer service

Published September 19, 2018

Optimise your customer service, please.

Good customer service is vital for good business. Everybody knows this, and yet so many businesses fail to prioritise it or continue to get it wrong. As businesses get more digital, a voice at the end of the phone or a message in a chatbox,  is often a customer’s first interaction with the brand. First impressions last.

As consumers we want the ‘human touch’ in our interactions with brands – we want to feel heard and understood, reassured that should anything go wrong, it will be empathetically dealt with and rectified on the other end through an exchange of social pleasantries, information and the desire to reach a solution on both sides.

Hello, is it help you’re looking for?

Automated responses remove this possibility. There are times when trying to resolve an issue with a delivery, for example, can leave us customers weeping into our tea in frustration and confusion at being processed through all the circles of hell with no resolution in sight. Passed from one senseless string of number options to another with no way out, we are left disliking the brand and solemnly vowing never to use them again.

Press #1 for intelligent service

It isn’t just automated systems that can make customers’ interactions with brands unbearable – human responders can be equally unsatisfactory but for different reasons. Human moods, emotional reactivity or sheer boredom on the job can produce the same negative results for brands with a ‘human help-desk’ whether that desk is in-house or in India. 

With rapid developments in natural language learning and AI there’s an opportunity to get customer service right, and relatively cheaply.  Until now companies have introduced technology in call centres or online in order to increase efficiencies and contain costs – but this has often been at the expense of customer satisfaction. Today’s AI has the potential to strike the perfect balance between human and automated services. Tomorrow’s AI, according to IBM, could replace humans altogether – with 85% of all customer interactions being handled without a human agent by 2020.

Press #2 for complaints

What are the pitfalls of AI replacing or augmenting either automated or human customer service systems? Much like the human teenage years, the growing-pain points for AI systems will likely consist of a combination of unresolved interactions, misunderstandings and giving in to peer pressure.

Unresolved interactions:

One of the most common complaints about automated services, even enhanced ones, is that the issue remains unresolved. This stems from the AI only being able to reach closure for predictable problems and/or there not being a sensitive enough algorithm or system in place that enables the AI to pass the complaint on to a human when necessary. Given the advances in Deep Learning and natural language learning – this issue should be resolved relatively quickly.

Misunderstandings:

As for unresolved interactions, misunderstandings that arise from the exceptions, subtleties and context of language can still lead down dead ends and rabbit holes. However, as was illustrated in the Google Duplex unveiling back in May, AI is evolving to be ever-more capable of contextualising and reframing around issues of accents, syntax and layered meanings.

Giving in to peer pressure:

Harmful human bias can pollute machine learning and result in chatbots that make racist assumptions, spout sexist slurs and generally reflect the hidden prejudices built into our language and our methods of collecting data and formulating decisions about the world.

“…For all their enormous potential, A.I.-powered systems have a dark side. Their decisions are only as good as the data that humans feed them. As their builders are learning, the data used to train deep-learning systems isn’t neutral. It can easily reflect the biases—conscious and unconscious—of the people who assemble it.” Fortune

Press #3 to speak to an operator

As deep learning progresses AI will ultimately  become indistinguishable from humans, especially in this area of customer service. The question isn’t if but when you should incorporate AI into your customer service strategy. In order to keep customers happy (or just to keep customers), brands will need to invest either in great AI service packages or in training and monitoring teams of humans. In financial terms, for larger businesses, it’s already an obvious choice to go artificial, but whatever the size of your business, incorporating AI that can take care of your customers in a meaningful way is worth exploring in both the short and long term.

Technology in marketing is nothing new, the trick is how to innovate your approach to win customer sentiment and grow your business. That is where tech marketing companies like ours prove useful: mad4digital is a tech marketing company for businesses that use technology or have developed technology. We position brands by focusing on what your technology can do and what problems it solves for the end users.

NEXT: AI, deep learning and targeted ads.