In her annual “megatrends” report, Kate Leggett, VP and Principal Analyst at Forrester Research for CRM and Customer Service, identifies three trends she sees in customer service in 2020. Her third prediction–that “agent staffing and culture will transform as AI sets in”–is interesting.
Though AI (really a form of automation) has existed in some form for many years, it’s only been recently that both have achieved levels of maturity permitting more widespread application and meaningful use. With the promise of greater efficiencies at lower cost, companies are racing to deploy these technologies across their business, including in customer service.
These revolutionary changes to the workforce have led to many divergent opinions. Most often raised is the likelihood of widespread job losses for those employed in routine jobs that can be automated and driven by AI (not limited to customer service). On a more positive note, customers seeking assistance will benefit from faster answers and customer service will be more critical than ever due to new technologies creating more questions for its users.
The truth is automation and AI are not new threats to the workforce; in fact, humans have sought to more speedily, accurately, and efficiently perform routine tasks for centuries. Despite these advancements, there is still plenty of work for humans today. As we have witnessed automation replace some forms of human labor, it has also raised demand for tasks requiring other skills and even creates new types of work. Again, nowhere is this truer than in customer service where humans aren’t going away any time soon.
Some customer questions and problems are more common than others. Be it how to set up recurring billing for a service to exchanging a product, multiple steps are often involved. These types of questions are typically higher volume and can be recited from memory by an agent.
Though easy for an agent to address, the level of AI commercially available is not yet at a point where it can reason through a problem it hasn’t been taught–even simple ones. While it excels at identifying the traits in the customer’s issue and recommending a likely solution based upon what it does know, it cannot solve what it’s unfamiliar with.
This is where humans assist. Machine learning (AI) analyzes volumes of cases closed by agents to “learn” how to identify a problem and recommend a solution. Customer service validates that understanding and adjusts it as needed. Only at that point can AI be used to direct both customers (and the agents assisting them) to likely solutions.
Connecting customers to answers
Analysts (including Leggett) have been saying for years that customers prefer self-service. When they have problems, they want to get answers at a time and place convenient to them. The most common forms of customer self-service in existence today are knowledge bases, chatbots, and online communities. Each of these to varying degrees rely on the existence of the human-documented solutions–knowledge articles–to be beneficial as well as humans to keep them functioning.
A knowledge base is a solution repository that continues to grow over time. Without proper curation, its value quickly diminishes: too many articles convolute searches, article quality declines as processes change, etc. Search logs must be monitored to understand under what circumstances customers succeed and fail to find answers. All of this care and feeding requires people.
Chatbots are the automation and AI solution to self-service. They also require their own maintenance. Humans determine the best issues for them to address on an ongoing basis. Chatbot conversations must be architected to quickly leads customers to the correct solution (which are often delivered in the form of knowledge articles). Terms customers use the ‘bot is unfamiliar are reviewed and added to its growing vocabulary.
Delivering complex and empathetic answers
To summarize, AI speeds problem identification from past cases and connects customers and agents to solutions while chatbots automate the process of delivering solutions. They aren’t applicable to every situation, though.
Not all issues are simple and have been documented. Some might be previously unseen and require questioning and decision-making to lead to the solution. The ability to reason through an issue remains a critical component to customer service, and it’s a skill only a human agent can provide today.
Some customer interactions require understanding and empathy. When orders don’t arrive on time, an additional fee is assessed, or a warranty doesn’t cover certain types of damage, customers won’t typically find the comfort they need and the resolution they desire from a machine. And though empathetic chatbots might be on the horizon, customers might not be ready to accept them.
What does the future hold?
We live in a moment in time where rapid change is happening all around us. Advancements in automation and AI seem to occur daily, and they are changing and improving the way the world works. This is especially true in customer service.
Automation and AI can speed the process of connecting customers to solutions. Yet without humans continuing to find and document the solutions, validating AI’s conclusions, and ensuring the automation is functioning, the technology would do more harm than good. As history has shown, technology will continue to change what work is in customer service, but it will be humans leading that transformation with new and different opportunities occurring in the wake of that change.