Can technology ever truly replace a present and attentive human mind?
It’s a question with philosophical undertones, but as Artificial Intelligence (AI) continues to evolve and surprise us, that isn’t stopping the tech industry from debating it.
While some in the tech world would have you believe that AI is on the brink of replacing vast swathes of the human workforce, now may be a good time to pause and think about just how much AI can realistically do on the ground level.
The ‘intelligence’ side of AI often captivates people more than the ‘artificial’ dimensions of the technology at hand. And while AI technologies by definition are capable of certain cognitive functions, they can only learn from the data put in front of them. New and unexpected scenarios can still stump the machines.
Humans, on the other hand, have the innate ability to adapt in real time, even in totally alien situations.
An example posed to CMSWire by Timo Elliott, global innovation evangelist at Walldorf, Germany-based SAP, illustrates this point:
“The modern world is full of complex but repetitive tasks that most of us would be happy to let a computer take over,” he said.
“A simple example in the finance department: if an invoice and payment match, the transaction can easily be processed automatically. But as soon as there are two invoices for a single payment, or the reference numbers don’t quite match, it takes a human being to sort out what’s gone wrong.”
These theoretical issues are compounded by the very raw problems AI is running into in the field. The vulgarity of the Microsoft Tay disaster springs to mind, while recent studies have exposed how AI programs can exhibit racial and gender biases. Once again, the simple fact that machines can only learn from what we serve up means they at times perpetuate the worst traits of humankind.
To get a firmer grasp on where AI technology is today, and whether or not we’re overestimating its practical usage, CMSWire spoke to some well-placed executives to gauge their perspectives.
Are businesses overestimating the practical powers of Artificial Intelligence?
Peggy Chen, CMO, SDL
After completing her Masters of Engineering in Electrical Engineering and Computer Science at the Massachusetts Institute of Technology, Chen started work at Oracle before joining SDL 14 years later. Tweet to @PBC88
When it comes to crunching data and automating mundane tasks, AI is incredibly beneficial. However, when it comes to customer interactions, machines still have a lot of learning to do.
Until they are able to more fluently emulate people, there should always be an integrated human touch readily available. Chatbots can be useful for answering quick and easy questions, but when customers are having a negative experience with a brand, only the most advanced chatbots are capable of detecting this negative sentiment and responding in an appropriate way. Consumers appreciate the self-service approach the digital world has enabled, but it’s important that they can always connect with an actual person when they want. With information available online everywhere, customers need a way to comprehend it all and want catered, personalized experiences.
Ironically, AI has actually become adept at delivering these more custom, personalized experiences, but machines can only do so much. For this reason, organizations should strive to humanize their digital experiences through AI, but always in tandem with the human touch.”
Cali Tran, President, Valassis Digital
As well as holding the position of President at Michigan-based Valassis Digital, a media delivery company, Tran is an investor with sales, business development and acquisition experience in high-growth potential technology companies. Tweet to @Valassis
While the concept of chatbots is not entirely new, we have only scratched the surface in terms of how they can be utilized for consumer engagement. The recent wave of innovation in artificial intelligence has brought chatbots to the forefront — supplementing job functions. Bots can increase employee efficiency and productivity while allowing companies to react quickly to consumer inquiries, ultimately improving the customer experience. They are not, however, meant to replace human interaction.
Consumers tend to favor self-service and chatbots can be a first point of contact, but in the case of an angry customer, chatbots aren’t necessarily meant to handle these issues independently. If necessary, the bot should have the capability to forward the consumer to the appropriate person immediately at any time during the experience. In addition, the bot may have already addressed many of the initial questions, which can help the representative solve the issue quicker.
The 24-hour, online and real-time assistance chatbots provide can remove friction between brands and consumers while allowing shoppers time to learn about a product or service on their own terms. To make a chatbot 'smarter,' and ensure it better meets consumer demands, bot language 'scripts' should be customized to the business and products or services they represent.
While it’s clear it will take time for companies to make chatbots as useful as possible, along the way they should be viewed as tools to help engage consumers and deliver value — not solve every issue. That’s where humans come in.
Aviad Abiri, VP, Portfolio Sales Enablement, NICE
Abiri has built an 18-year career at London-based NICE Systems, the globally recognized customer experience and financial security firm. In his current position as Vice President, Portfolio Sales Enablement, Abiri takes control of ensuring that all client-facing employees have the capability to consistently conduct productive conversations with current and potential clients. Tweet to @NICELtd
In today’s digital age, customer service is not always easy. Customers are interacting with organizations on a variety of channels (surveys, social media, text, phone, etc.) and expecting immediate, personalized responses. While every single interaction is an opportunity for companies to connect with the customer, the millions of individual interactions can feel extremely daunting for service providers.
By using technology like AI and machine learning, customer service agents can better understand customer requests while optimizing their business processes. There will always be a need for a human interaction — especially for tricky customer service calls — but AI technology can help companies respond to customers with real-time, intelligent, meaningful interactions. Additionally, with advanced technology like robotic automation, organizations can also receive assistance with back office and reporting tasks, allowing more time for human-to-human interaction.
Humans will never be replaced, but machine learning will help augment and optimize a customer service agent’s day-to-day tasks.
Daisy Hernandez, Global VP of Product Management, SAP Jam
Daisy Hernandez is VP of Product Management for SAP Jam, SAP’s social collaboration cloud product. She is responsible for driving the product vision to solve business challenges by facilitating meaningful interactions between employees, customers and partners. Prior to SAP, Daisy held several leadership roles in business operations, engineering program management and software development at companies such as Oracle. Tweet to @mmcHernandez
Customers and vendors alike are still identifying the best ways to apply AI and chatbots for the right scenarios, both for internal and external use. There will be adjustments to how AI is being applied based on lessons learned, which naturally happens with most cutting-edge technologies. There are certainly some cases where using a chatbot to interface with a virtual assistant will be useful, and many others where it will be inappropriate or harmful. Whether chatbots are an asset or a liability depends heavily on what the person needs and how simple or complex their request is.
For example, if a customer knows exactly what they want, and it’s a straightforward and simple request — such as getting the status of a flight or delivery — then the expediency and simplicity of a chatbot will be much preferred by most customers. There will certainly be other cases, though, where resolving an issue will require an actual person.
This is no different than being shuttled to a 'phone tree' system when you dial the support line for a product or service. How many times have you been frustrated when a phone tree doesn’t give you a simple option to talk to a person right away? If you as the customer already know your issue will require more complex interactions than just punching in codes, being pushed off to an automated solution will most definitely become irritating and time consuming. The key to all of this will be dependent on whether chatbots and AI are developed to understand intent, need and complexity.
Rajeev Shrivastava, Chief Strategy Officer, inContact
After Co-Founding VirtualSoft Systems in 1998, Shrivastava went on to work for tech giants like Oracle and Rackspace. As CSO at inContact, he oversees the teams responsible for overall strategy, product management, user experience, partnerships, business development and M&A. Tweet to @Rajeev_Shri
While many organizations are currently using AI in meaningful ways, it’s definitely not the solution to every problem. In the customer service industry, specifically, there’s a lot of promise for what the technology can do to improve overall customer satisfaction. With AI supporting routine queries, customer service agents can focus on more complex interactions that drive customer satisfaction. Organizations are looking to Chat bots and speech recognition technology to automate routine service interactions, drive enhanced agent productivity and thus improved customer satisfaction.
Some think the future is AI in order to deal with the growing number of customer interactions across a multitude of channels, companies must not forget the human element. The answer is in integrating AI with traditional, conversational communications customer experience. There is no replacement for empathy, and human interaction will always be a key element for a positive customer experience in the contact center.