Every year the latest technology makes our lives faster, more mobile, and a little less human. And the same can be said for business. As the proliferation of channels, data, and innovation make transactions more complicated, businesses tend to look towards automation as an answer to the modern-day customer woes. And it’s becoming quite the reality, Gartner predicts that by 2021, artificial intelligence (AI) spending may exceed £20.9 billion.
As I’ve discussed previously, the latest advances in AI customer service and machine learning (ML), however impressive, do not justify a customer care approach that relies heavily on virtual associates. Automation is not a silver bullet and companies need a thoughtful approach to managing AI customer service and ML. Remember, customers still want to talk to humans and companies still want to talk to their customers.
I recently had the opportunity to speak at the CCW trade show in Berlin, where I spoke with some of the brightest minds in customer care. We discussed how a balance of AI and humans can create an amazing customer experience.
What AI customer service does well
AI customer service has a wide array of tools that can vastly improve your businesses customer experience. As customers continue to demand services that are omnichannel, 24/7, and personalised, it’s hard to find the right associates, specially when you compete against highly disruptive companies like Netflix and Amazon that have defied convention and raised the customer experience bar. You need to place yourself in your customers’ shoes - if they can’t have a service exactly when or how they want it, do they actually need you?
That’s where automation comes into play. Your businesses’ bots can be available 24/7 to put out fast, and consistent responses across channels. In addition, it reduces costs by scaling without hiring more associates. While at the same time, deflecting calls enables associates to take care of high priority calls or more complex tasks. AI customer service can handle surge volumes by answering simple questions on delivery statuses, complaint resolution, balance enquiries, product searching and more. Overall, Juniper Research predicts that virtual assistants will help industries save up to £5.97 billion per annum by 2020.
Where AI customer service can frustrate
AI customer service works great ‘if’ it operates correctly. More than often the quality of its natural language understanding (NLU) is weak, it’s easy to stump the highly scripted systems with casual language, and many can’t escalate to a human associate properly, prompting the customer to repeat their question.
Even Facebook, one of the world’s most popular media outlets, experienced a 70 percent failure rate with its bots in 2017 according to the Motley Fool. The article attributed this as one of the consequences of AI that cannot handle complex enquiries, which leads to a frustrating customer experience.
In addition, since bots aren’t connected to key systems, it has very limited personalisation, which is a real deal breaker for a society that’s grown to expect plenty of personal touches in their shopping experience. Especially if we keep up with general trend of moving towards AI. Last year, Gartner predicted that by 2019 over 10 percent of IT hires may be writing scripts for bots in customer service.
A perfect partnership
For all its challenges the time to invest in AI customer service is now. Research from Ubisend found that 69 percent of customers would talk to a bot before a human if it can provide instant answers, but this is only to a point. Current state shows 67 percent of customers would rather speak with human associates on the phone or chat over automation according to a NICE inContact report. The solution: a hybrid service workforce where the bot simplifies, and the human associate engages.
First, this means investing in your AI ML to transform from ‘dumb bots’ into intelligent virtual associates (IVA). A better understanding in NLU and ML turns your highly scripted and channel specific bots into adaptable, multi-channel AI associates. This virtual assistant can respond with personalised and contextual answers, remember what it has learned, and when necessary seamlessly transfer to a human associate.
These advanced tactics can be trained by using the same material one would use for live associates, structured into a format it can understand in the training platform. Then only when its proven it can handle interactions with live associates will it be released to the customer. This will shrink the amount of shoddy bots that crash at the slightest inconvenience.
When you have associates and bots learning and working together in the same environment, business can see a symbiotic relationship unfold. The live associates can enforce quality control over the virtual assistants on new topics, and in return the virtual associates can assist human associates in answer retrieval.
A chatbot relay
Let’s look at how this ideal relationship may work. A customer initially reaches out to a bot on your businesses app to look for camera supplies. The smart virtual assistant would recognise the customer using the email he or she submitted to initiate the chat. Then, using its integrated knowledgebase it can leverage product history and suggest what the customer may need (with links and pictures to make it clearer). If the conversation isn’t going the customer’s way or it doesn’t have the access needed to answer the question, it can use instant channel migration to transfer the customer to a channel of choice, in this case a live human associate.
Once in the hands of the associate the omnichannel platform can perform a smart hand-off, transferring all the customer’s interaction data to the real person. This change of hands is strengthened by intelligent routing, in which the customer is matched with an associate who shares their interests, thus creating empathetic and proactive experience that builds loyalty. In this type of situation both the virtual and human associate’s training will allow for operation excellence.
Taking responsibility
Bots will never replace the personalised care a person can offer, and neither will we ever access information in a nanosecond. However, the point of any relationship is the ability to thrive off each other’s strengths. If businesses take the time to not only improve the quality of AI customer service before it’s released, but also take a second look at the role it plays with humans, we will see a new Golden Era in customer experience.
