In my day-to-day work partnering with leading brands and retailers to solve customer challenges, I’ve often heard the statement, “We are behind.” Over time, it's become clear to me that they aren't so much referring to being behind their competitors, but rather that they are falling behind customer expectations.
For many businesses, chasing after the customers’ ever-changing behavior and expectations has led to a whole new focus of business strategy, where services and customer experiences are now the battleground for customer loyalty. Increasingly, as Forrester and other industry analyst firms predict, this means developing digital-first customer experiences that leverage the power of customer care automation to provide the highly contextual, artificial intelligence-informed pre- and post-purchase assistance today's customers demand.
Gartner notes that by 2020, 25% of customer service will feature intelligent automated assistants. For technology and business leaders looking to take advantage of the revenue-generation, cost-saving and relationship-building benefits that customer care automation offers, there are six hidden challenges you'll need to address that I have experienced in my own line of work.
1. Information Lookup
While the majority of customer service inquiries may not be complex, they do require more than a simple information lookup to get to a resolution that satisfies the customer. Most inquiries require going through a business-specific workflow and eventually taking action on behalf of the customer -- initiating a return, helping them update their shipping info, checking their loyalty program points, etc.
Knowledge-based technologies, such as FAQ tools, are not solely able to create a sufficient level of automation to fully support most customer service requests. Decision-tree-based solutions are often needed to map out service workflows; however, they come with a cost -- high overheads in crafting the dialogue possibilities that fulfill the workflows. Many of the innovation opportunities lie in automating dialogue generation and optimizing service interactions.
2. Access To Data
The speed of customer service inquiry resolution hinges on the ability to access, aggregate and apply a particular customer's relevant data. A simple support case regarding a customer’s past purchase requires a wide array of information, including data stored in third-party systems and other platforms.
Not only does your automation solution need to integrate with such systems, but it also needs to be able to consume the data from them at real-time decision-making speeds. Customer-relationship management (CRM) systems are typically built without real-time decision-making applications in mind, so integration alone may not solve the data access that's required.
3. Contextual Understanding
Customers expect personal service and fast resolution without hassles. No one wants to communicate with a customer service agent, human or automated platform that seems to have the memory span of a goldfish or doesn’t recognize them or their needs.
Contextual-understanding-enabled inference capability coupled with natural language understanding (NLU) can allow an AI-powered automated assistant to discern a customer’s “unspoken intent,” reducing unnecessary back-and-forth and increasing the resolution speed as well as the quality of the answer.
When building an assistant with a versatile skill set (i.e., the scope of services the assistant can offer), the number of possible dialogue flow scenarios increases exponentially in relation to the number of services the assistant offers. From an implementation perspective, as the skill set grows, mapping out the complete natural dialogue flow quickly becomes overwhelming.
Look for an automation solution that enables your different teams to “teach” the assistant new skills without having to possess deep knowledge about what skills it already has or how they work. These different skills and services should work in harmony, instead of it being a problem for your team(s) to address internally, and it should be seamless for customers.
Customer care automation isn't a zero-sum game. There is still a well-defined need for human agents, and bot-agent collaboration should be part of any sound automation strategy.
For most businesses, automation is a road map for going from a small number of automated services to a large number of services. An effective bot-to-agent handover demands more than just integration with agent-facing systems; it can allow the automation solution to provide agents with augmented information (i.e., context) that can improve agent performance. So, make sure to map out a solid solution for the bot-to-agent handover when building your customer care automation strategy.
6. Channel Experience
The rise of conversational channels like voice platforms (Amazon Alexa and Google Assistant, for example) and messaging apps like Facebook Messenger poses new challenges for automation solutions, which need to deliver seamless cross-channel experiences. Voice platforms, especially, create a whole new set of user experience (UX) requirements with strong dependencies on AI capability. And as modern messaging apps are asynchronous chat applications, even when the topics change, the conversation (at least on the customer’s end) is a single blended thread.
A channel-agnostic automated assistant solution needs to understand the nuances of these differentiated user experiences and possess the ability to use each channel (voice, chat, SMS, etc.) to the best of its possibilities.
Launching Effective Automated Customer Care
The opportunity to deliver effective service, engagement and even sales through an automated assistant is a very real one, with a multitude of benefits that make taking action not only a wise approach, but one that will dramatically separate those who do from those who don’t -- much like mobile web did for businesses five years ago and other customer experience innovations have done before that.
Leaders across brands need to recognize that delivering automated experiences customers love is not a trivial task, and it requires challenging problems to be solved, but the rewards are well worth the effort.
This article was originally published on Forbes.com