Shifting work to your customers is an oddly attractive way of driving efficiency. It is often called self-service, but that term can be an oxymoron. It is, of course, efficient for me to force you to do work that I used to do, but it is clearly inefficient for you. And yet this shifting of work has become frustratingly common – and it brings with it costs that companies often fail to consider in adopting automation.
Well-designed self-service can be very effective, but the terms ‘self’ and ‘service’ go together only when the customer’s needs are at the center of the design. Anything else is just sloughing off work on another. It is shifting work, for example, to ask a customer to (try to) solve their problem on the web (unless this is his or her preference). It is clearly not self-service, but the act of shifting work, to put a customer on hold for minutes while you wait for the 'next available agent'. The menu of options that customers are often forced to navigate to get service, are also not really self-service; they are organized according to the company’s internal logic of operations, not with the customer’s concerns in mind.
Today’s technology is AI. Although an intelligent bot can be helpful for some routine queries, it often comes without an easy escape hatch to a human being. It is not self-service when reaching a human representative is like running a gauntlet. The fact that bots are often deemed by their creators to be intelligent does not help when, as is frequently the case, they do not work. Around half of surveyed Americans say the increased use of artificial intelligence in daily life makes them feel more concerned than excited, according to a 2023 study by the Pew Research Center.
Each of these forms of shifting work depends on introducing a new technology, from call distributors and interactive voice response systems to voice recognition and generative AI. The driving goal almost always starts as a grasp for efficiency, almost never with customer satisfaction.
This is myopic automation.
There are, of course, exceptions. Amazon is a good example of a company that is relentlessly efficient but focuses on the customer experience first. The overriding priority of Amazon from its inception has been customer obsession, across all functions.
"When you’re obsessed about something, it means it’s blinding to you; it possesses every ounce of your existence and being,” according to John Rossman, who helped start the Marketplaces business at Amazon. In ‘Innovation the Amazon Way’, published in Research-Technology Management, he explains that “Everybody is expected to understand the customer deeply. You need to understand customers not just at the narrow intersection at which you happen to be dealing with them, but in a broader sense as well."
This includes those introducing new technology to automate work.
Amazon self-service systems usually just work, which is ideal for both the customer and the company, but Amazon backs them up with an equally important capability: when something goes wrong, you can reach a person who is empowered to solve your problem – even if it is your fault and will cost the company money. That empowerment, ironically, is essential to good automation because it provides the resilience needed for effective support.
ASAPP is another example of a company that focuses on customer service first and business outcomes as a consequence. It is an AI company that makes software for contact centers. Joe Ciuffo, ASAPP’s Head of Marketing, told me in an interview that companies often try to set up a “balance of what's good for the brand and what's good for the customer, and it becomes a tug of war.” In times of stress, the customer always loses.
ASAPP has decided that the focus in creating good service should be to help agents to help customers. That has meant using technology to get away from scripted dialogs and to enable more natural conversations with the customers. “Conversation first, keyboard second” is how Ciuffo described it to me. When this natural dialogue is combined with agents who are empowered to make decisions that matter to customers, business results follow.
What are the business results of designing work for service? The first is happier customers, which can lead to improved customer retention and higher average sales per transaction – and even to lower return rates of products. A second benefit is reduced turnover of agents, with all the costs associated with hiring and training new representatives. In addition to Net Promoter Scores, which measure customer satisfaction, ASAPP customers often use Employee Net Promoter Scores to measure associate satisfaction, which is correlated with reduced turnover.
How do companies achieve automation and good customer service? The prerequisite is customer orientation: customer first, worker second, with business results following as a consequence of good service. This can be difficult to sustain in a pressured business environment, but there are approaches that help.
Connect the technologists with the work: ASAPP develops its software in collaboration with the rep. The technology is designed to adapt to agents' needs, not the other way around. This is accomplished with engineers who spend time with the reps and by enabling those on the front line to suggest new features that will help improve their jobs – and by extension, customer service.
Empower your agents: Amazon empowers its agents to refund customers related to product delivery. They monitor customer behavior to prevent abuse, but they start with the premise that the customer is honest.
Broaden your metrics: Customer-centered companies broaden their metrics for customer service to include both customer and worker-satisfaction. They also look at hidden costs, like repeat callers and turnover.
Beware of efficiency promises: Vendors may promise extraordinary savings, but they often fail to take into account hidden costs. Today, there are companies that claim that their generative AI can eliminate the vast majority of customer service agents. Starting with a goal of eliminating positions is putting the cart before the horse and is unlikely to be successful. Make sure that analysis of the automation includes all costs.
What can you do if you are an executive sponsoring the introduction of a new customer support system? How can you assure that the resulting work will be not only efficient but effective? Start by asking questions. Ask about the customer experience that the company is trying to create. Be persistent and press for details. Ask how customers can bail out to reach a person, if necessary. Ask how we know it will make a customer’s life better. Avoid a focus on efficiency metrics.
Another suggestion is to become a user of your own customer service. Experience it yourself, first-hand, and compare it with great experiences that you have had with other companies, even those outside your industry. Would you return to your company based on the experience? Go further, and listen to a few calls; see how customers react. Ask your own employees how they experience doing the work that sustains customers. Don’t do this once, but make it a practice.
Finally, seek to surface hidden costs. The benefits of automation are often calculated directly in terms of the number of calls eliminated and the number of jobs reduced as a result. But things rarely happen the way you would want them to. Try to identify what could go wrong with the customer experience as you implement any new system. During implementation, pay special attention to metrics of customer and worker satisfaction. Make these questions part of every review.
It is possible to have both efficient work and good customer service. Getting there starts with a focus on the customer, not the magic of technology. True productivity will follow good design.
