COVID-19 has been jarring to all facets of business, but perhaps no function has been more disrupted than customer support. Customer service is the frontline for any business — and right now, there are still a lot of anxious consumers that need help.
In many industries — especially those like hospitality, travel, and entertainment, to name a few — businesses have seen incredible spikes in customer outreach. Meanwhile, the spread of COVID-19 has forced many brick-and-mortar contact centers to close. Companies have rushed to mobilize call-center agents — along with the technology they need — to work at home. But many agents simply don’t have the appropriate environment to work at home, especially if they have children and extended family also sheltering-in-place.
More customers need help. Fewer agents are able to provide that help. That’s a recipe for phone support queues several hours long and insurmountable backlogs of tickets that no contact center — let alone one that’s physically closed — could deal with.
As consumers, most of us have been more tolerant of slow customer support — but that patience isn’t endless. As my co-panelist Paul Brandt (Pizza Hut’s VP of customer experience) said in a recent webcast, “Consumers may be more forgiving, but they are still looking to companies for help. And they’ll remember who was there and who wasn’t.”
Efficiency alone isn’t enough. We need resilience
For years, customer service organizations and leadership have almost singularly been focused on improving efficiency. Our human systems are optimized to be more efficient. Technology, we’ve opined, is a means of adding scale so that our support teams could be more efficient — to handle more issues. More volume. More customers. More growth. Less cost.
The problem with designing efficient systems for today is that they aren’t necessarily efficient tomorrow. COVID-19 has illuminated the difference between static systems that are efficient today and dynamic ones that are flexible and intelligent enough to respond to sudden changes in the market.
Much of the legacy automation we experience in the form of IVR systems and early chatbots are specifically designed to efficiently handle the top issues looking back. They may be intelligent in that their design was fueled by data from the past. But they aren’t connected to what’s happening now, nor are they going to respond to unexpected events tomorrow. Even without a crisis, these disconnected systems age fast as companies and industries naturally evolve — and, in parallel, the customer experience worsens.
These static systems are historians. They’re not futurists.
They will not be resilient tomorrow. They won’t be able to account for market disruptions, or even disruptions in your own company, that no one can predict. They won’t help when a call center is shuttered or a surge of new issues surfaces related to issues that no one could have anticipated. At least not without a lot of human intervention.
Efficiency alone isn’t enough. We need customer service systems to understand what’s happening now, to be intelligent in the moment, to be resilient. Resilient, like our people.
What resilient customer service looks like
So, how do we create customer service today for a tomorrow that we can’t possibly predict? Ultimately, it’s the difference between a system that’s static and one that’s dynamic. It has very few fixed pieces. It has a flexible location. It can scale up and down as needed. It includes redundancy so that if parts of a system fail, others can cover. Here are the key ingredients:
AI systems that understand trends before we do
Many companies have teams of analysts pouring over customer service logs on a daily or weekly basis to identify new issues. They produce reports of top issues and present them to executives. These reports become the basis for creating new content to answer questions or building automation. It’s an incredibly static, and typically very slow, process.
Humans on their own simply aren’t adept at quickly identifying new issues — but AI systems with humans in the loop can be. Machine learning can analyze millions of conversations in a matter of seconds. It can identify relationships between issues that we humans simply can’t. If these systems tap knowledgeable humans to interpret and review the results, magic happens. The AI system can get so good that it can even help predict tomorrow’s disruption based on what’s happening today. AI can help us understand.
To have a dynamic AI system that understands what’s happening in the moment means connecting it to business systems and domain experts. AI systems will need smart information architecture. Only then can the algorithms monitor every call log, every chat session, every email as it happens. This AI must also use more sophisticated natural language processing to understand emerging issues and intents — and account for communication nuances on different channels — rather than simply recognizing the top issues from yesterday. CX leaders won’t need to wait for a weekly or monthly report —these AI systems will tap the right information and the right people in the loop — so CX leaders can know in the moment when a new issue is emerging.
Dynamic and specialized knowledge bases
To ultimately serve up great customer support requires knowledge. Whether served up by human agents, virtual agents, or some other digital tool, answers need to be dynamic for customer service to be resilient.
The knowledge base of the past, typically created by internal support teams and pushed out in the form of FAQs and self-service websites, isn’t nearly resilient enough. These internal-led resources often take weeks or months to create and push through approvals. They may generally address broad topics, but seldom are they granular enough to effectively answer specific questions. There’s simply no way for teams to plan ahead and create an answer for tomorrow when they can’t anticipate the question.
To be truly dynamic, we need a network or community of expertise and a system that can harness its collective intelligence. Tapping freelance workers with specialized expertise can give companies the diversity of answers they need for today’s issues. It also provides scale, enabling faster responses than a finite team can provide when an unexpected disruption emerges.
Companies and customer support teams have been investing in process automation for years to try and become more efficient. But in many cases, this automation is disconnected from support systems — and is designed to automate top systems based on past reports. These static implementations often come in the form of IVR systems that force callers to choose between a fixed number of selections. Sometimes, it’s a chatbot that uses similar “decision-tree” logic.
The problem with these automations is they don’t reflect what’s actually happening in real-time. They were one-off projects that were designed and developed, but not meant to be continuously updated. They are anything but resilient.
To be resilient, we need intelligent automation that is agile — and a system that allows companies to easily respond in a market disruption. This system should include virtual levers that allow leaders to adjust volumes for different channels by diverting certain issues and automating other ones — for example, an overwhelmed IVR system could push callers to digital messaging platforms. Also, companies should leverage machine-learning predictive routing that directs customers to a resource most likely to be able to answer a specific question — whether it’s an AI-powered chatbot, an external community expert, or an in-house support agent.
Decentralized contact centers
The idea of a geographic “contact center” was already losing its relevance when COVID-19 suddenly forced many brick-and-mortar centers to shut their doors. Some companies had geographically diversified centers and teams — but in a global pandemic, there’s no amount of geographic diversification that helps.
To make matters worse, many of these centers rely upon static hardware technology that is difficult to mobilize. From desktop computers and custom communications devices to the servers that power custom applications, these physical centers are incredibly static. As one analyst firm described, this “tired butts-in-seats delivery model” was exposed by COVID-19, and things will need to change.
The customer service industry needs to follow (finally) other knowledge worker professions by supporting more agents to work remotely and at home. This will require an update to the “call center” toolkit with hardware and software that support a distributed agent model.
Companies also need a more dynamic approach to its pool of available human agents — a fixed number of full-time agents simply can’t respond during surges caused by a market disruption. To be resilient, companies need “on-call” agents — which could be flexible part-time workers and/or a gig workforce that is paid based on resolving individual cases.
Resilience may be hard to measure. But it’s our new imperative.
Customer service has always been very metric driven. It will continue to be, as it should. We need ways to measure and improve performance of our customer service teams and systems. We need data to optimize the performance of AI and virtual agents. We’ll continue to look at metrics like CSAT (customer satisfaction), resolution rate, cost per resolution. These metrics often comprise a business case companies would use to build new customer service solutions — and then help them measure ROI after new technology is deployed. These are measures of efficiency — and they allow us to look back and see how we performed.
However, as businesses begin to re-open, we need to look ahead. We need a plan for how we can be more resilient, in customer service and beyond.
Resilience is harder to quantify in a spreadsheet. Being able to respond to the unexpected is difficult to measure. After all, there’s no benchmark for COVID-19. Our industry has never seen anything like it.
Even if it’s difficult to put a number to it, we know that we must be better for the next crisis or market disruption. We need to be there for our customers when they need us most.
That’s why resilience is our new imperative at Directly. We need customer service systems to be resilient.
COVID-19 has forced many in our industry to re-think how we provide customer support. Companies such as Microsoft, Samsung, Airbnb, and Autodesk are using Directly’s platform to deliver more automated and elastic customer support operations that are resilient, even in the face of a crisis. Our platform integrates with support channels to understand customer issues, automate common solutions, and engage community experts. This enables customer support leaders to resolve customer issues with the right mix of automation and human support, boosting customer satisfaction, while saving millions per year. Contact us to set up a demo of our platform today.