When we think about the most impactful technologies in the last decade, one can argue that mobile and cloud have made the most impact on customer service (CS). The explosion of mobile phones was not just about a change in form-factor, but the emergence of the app economy has transformed expectations in the customer service experience. The growth of messaging and social apps created a sense of urgency towards customer service and a network effect with a need to share positive and negative experiences.
The growth of cloud as a technology became the foundation of omnichannel experiences wherein customers engaged with a brand from a channel of their preference. Cloud also introduced fast-changing enterprise software delivered as a service (SaaS) for customer relationship and service management. The myriad of applications in the ecosystem for customer service management dramatically changed the process of managing customer service. These systems have now become the systems of record for customer service.
Digital transformation for CS leaders means they have to build a single view of the customer across channels, issues, processes, etc. In addition, they are challenged to offer a consistent experience across channels and traverse context as consumers move across channels. Businesses are innovating at a rapid pace, aided by cloud and mobile to address these needs.
CS leaders, faced with increasing internal pressures and accountability, need to prioritize and improve digital experiences, and explore digital applications to improve operational efficiency. This has increased the complexity of CS leaders’ portfolio and has, at the same time, made the customer service function strategic.
As we look at the next decade, here are some of the trends and challenges that CS leaders have to plan for:
i. Understanding the 360-degree view of customer interactions
Customers’ shift to omnichannel support is significant where they simultaneously interact with the brand through multiple channels. Context is transferred not only across channels but through all the interactions a customer has had with your company across sales, marketing, and success. Many customers are inclined towards using a mix of social channels along with email or voice support to get their voice heard with the brand.
ii. Explosion in customer service data
Data related to customers and their service needs is exploding. This includes system-of-record data in CxM systems, content for customer service, data on social channels, environmental data related to business’ products and service, etc. When analyzed appropriately, the insights from these data sources can be converted to systems-of-engagement and systems-of-actions.
iii. AI to become mainstream in customer service
While the last decade has been about trying to find the applications of AI in customer service, it is fast becoming mainstream as a key productivity enhancer for companies worldwide. Gartner reported that 55% of established companies either have started making investments in the potential of artificial intelligence or are planning to do so by 2020.
iv. Rise of intelligent process automation (IPA)
Intelligent Process Automation (IPA) as a concept is the technology for CS leaders to embrace to realize and maximize productivity gains and cost savings. Intelligent Process Automation (IPA) uses Robotic Process Automation (RPA) with the cognitive aspects of AI to automate mundane and repeatable customer service processes. This allows customer service agents to focus on making more human and empathetic conversations with customers.
v. Single channel to multi-channel to omnichannel
With the growth in the number of channels for customer support coupled with the changes in customer expectations, CS leaders now must plan to provide a seamless and integrated experience across all channels simultaneously instead of using various channels independently.
vi. Gig economy in customer service
Companies like Uber, Airbnb, DoorDash, etc. have led to income generation by monetizing any free time, space, and skills, creating an army of freelance professionals. CS leaders have a challenging set of problems associated with the immediacy, high volume and the cost of supporting real-time expectations.
vii. Growth of social as a channel
Social as a channel has emerged with greater prominence and will be tricky to manage as brands gain a lot of influence and advocacy using the network effect. Similarly, a complaint on social media like a tweet or Facebook post has everyone’s attention due to the immediacy and the network effect.
In light of these trends and challenges, it is important that CS leaders understand how to get ahead in the next decade and use technology as a way to transform customer service and their careers.
So, what does this mean for CS leaders?
The two most transformative and impactful technologies are AI & Intelligent Process Automation (IPA). Let’s start by describing what these mean in the context of customer service needs and their impact in the coming decade.
Artificial Intelligence (AI)
AI is simply the concept of machines learning from past interactions, data & results and mimicking human actions and decisions. As it relates to customer service, it is easy to imagine some customer needs that machines can ‘learn’ from existing data and improve with time. Also, the cognitive output from AI can be used to automate repetitive and mundane tasks that a machine can be trained to mimic using RPA.
Intelligent Process Automation (IPA)
Businesses driving automation in their processes frequently run into buzzwords like Robotic Process Automation (RPA) and Intelligent Process Automation (IPA). RPA refers to software that can be programmed to do basic tasks across applications and systems just as human workers do. The software robot can be ‘taught’ simple repetitive workflows representing processes with multiple steps and applications, such as taking received forms, sending a receipt message, checking the form for completeness, filing the form in a folder and updating a spreadsheet, and so on.
Intelligent Process Automation (IPA) is a combination of AI and Robotic Process Automation (RPA). It essentially is a software that mimics the behavior of a user and is being used in mainstream business scenarios, primarily for process automation. The range of applications could be from collecting and combining simple data and making contextual decisions to drive a process, to providing automated responses from within a process.
The cognitive aspect of AI is used to understand the intent that triggers the execution of processes and also to contextually provide templated responses at different phases of the process. This type of automation is ideal for several customer service scenarios like refunds, warranty, exchange, order processing, tracking, invoicing, payment update, inventory look up, etc.
Every business has processes that the customer service professional has to perform to resolve cases. There are some repeated customer service needs that require follow-up through processes touching multiple systems, decision points, and approvals. These are particularly labor-intensive, such as new business applications handling, policy change administration, claims set-up and various finance and accounting activities. Intelligent process automation (IPA) of these repeated tasks reduces reliance on multiple systems, reduces errors and improves productivity, efficiency and effectiveness of customer services.
The potential for using AI & Intelligent Process Automation (IPA) in customer service is huge but the usage of these technologies has been low till date. We must ask – why has this been the case? The problem lies in the fact that CS leaders have been bombarded with buzzwords like Deep Learning, Neural Networks, Natural Language Processing, Chatbot, Virtual Assistants, etc. that has the CS leaders confused and fearful about where to start and how to go about it.
Where do CS leaders start?
To successfully create a concrete plan and implement AI & Intelligent Process Automation (IPA) in customer service, the logical place to start is across these 2 dimensions:
1. Understanding the mix of product type & channel for implementing AI-based offerings
2. Creating a business case for AI-based solutions
Let us discuss both options in further detail below.
1. Mix of product type & channel for implementing AI-based offerings
CS leaders need to first pick the optimal channel where they can implement AI between Voice, Chat, and Email. The advancement and success of AI technologies across product types and these channels differ and it’s important to understand where your organization can get the best ROI.
Let us look at some types of AI-based product offerings and what channels are applicable and best suited for them below.
A. Self-Service: It is said that the best form of customer service is no customer service. CS leaders must, therefore, focus on the help center as the first point of entry for self-service. Self-service technologies that serve digital channels across web, email, and chat can reduce incoming case volumes by as much as 20-30% if implemented well across all channels. Implementing AI-based solutions in the help center can deliver a high ROI as a significant number of these cases can be resolved at a fraction of the cost. Self-service AI can be implemented in two ways – as a submission form or as a conversational form.
I. Submission Forms: These are the most widely-used help center interfaces for AI-based solutions that can deflect informational cases and cases that need simple processes. Informational cases are resolved using articles from a knowledge base.
II. Virtual Assistants: Virtual assistants or Chatbots, the biggest hype of AI in recent years is easy to build but it’s expensive and hard to make it learn and improve over time. CS leaders should be wary of systems integrators that offer an inexpensive chatbot to start the automation – Most system integrators don’t have a complete understanding of the issue heatmap, decision trees and the type of content in short form that resolves a case.
Self-service is largely driven using content in the knowledge base. Here the customer writes his/her query in a submission form or a chat interface and the AI engine understands the content and the context of the problem and uses the articles in the knowledge base to recommend top articles that match the user’s problem. The customer is able to solve the problem using self-service and no ticket is created.
However, to build, maintain and ‘learn’ over time, virtual assistants as a category of chatbots that are conversational in nature need to be designed to respond to customer questions using knowledge articles as well as assist in automating processes. These virtual assistants should learn from ingesting historical cases/chat history, have a good understanding of support heatmap and should help design decision trees to resolve complex processes.
B. Automated Triage: Case classification is the first place to implement AI within any large organization with several products or services. With the advancement of Natural Language Processing (NLP), one of the many AI technologies, there shouldn’t be any need for humans to perform triage across any channel. This includes:
- I. Voice: Speech recognition technology has reached amazing levels of accuracy and can be experienced in the way Siri, Alexa, Google Voice, and other assistants perform. Modern IVR platforms are exceptional at understanding phrases, accents, and nuances, to transfer a call to the right department. We’re also seeing businesses ask their customers to leave a voicemail and using AI-based speech to text transcription to call back with a precise understanding of the customer’s issue.
- II. Chat: The hype of chatbots is something you’ve all experienced in the last few years as being the best thing since sliced bread. We must reorient our thinking towards using chatbots to primarily focus on triaging and solving simple, known issues.
- III. Email: This is the channel where AI-based triage is most advanced and one where you should never have any humans triaging a case. Advances in NLP have given the ability to read a paragraph and understand the case type, category, sub-category, customer tier, priority, sentiment, language and a lot more.
C. Automated Responses: The email channel is perfect for automating responses using AI once a case has been created. Email templates or macros are sure-shot ways to resolve cases without agent interaction. CS leaders should move away from rules and triggers-based auto responses to using AI-based auto responses that leverage natural language understanding.
A common form of implementation of auto response is from within a CxM system. Here the AI engine is able to understand the intent of the customer problem and then use templated responses based on historical replies from agents to answer the customer’s problems. The ticket is auto-replied to without any agent touching it.
D. Intelligent Process Automation (IPA): Since Intelligent Process Automation (IPA) is a combination of AI & RPA, it uses the cognitive power of AI to understand the intent of customer query and hence the email or the chat channel is the most obvious channel to start with for Intelligent Process Automation (IPA).
Here the AI engine is able to understand the intent of the customer problem and then map it to an internal process which is automated using RPA. The process is triggered and completes automatically to resolve the customer problem and the ticket never touches an agent.
E. Agent Assistance: Canned responses and knowledge articles are two of the most useful resource types that boost agent productivity. Here the AI engine is able to understand the intent of the customer query and suggest recommended knowledge articles and templated responses that can be used by the agent to resolve the case.
2. Creating a business case for AI-based automation using zero contact resolution
Creating a business case for AI-based automation starts with defining and measuring success. CS leaders are always under cost pressures to do more with less, use technology to manage seasonal service volume spikes while maintaining the morale of the agents. The question is what AI-based solutions can do to drive the customer service function from being a cost to adding value to businesses. CS leaders have three core metrics, in addition to others that are in their control: Average handle time (AHT), First contact resolution (FCR) and agent satisfaction. However, the businesses they support measure the CS function using business metrics like CSAT, NPS, LTV that are often a result of the core metrics being healthy. CS leaders need to rethink traditional CS metrics and think about adding an automation-relevant metric like Zero Contact Resolution (ZCR).
ZCR is the concept of resolving cases without agent interactions. ZCR is a core AI metric that smart organizations leveraging AI can take advantage of. The conversation can move from just being a cost center to value delivered by leveraging AI technology to reduce the case volumes that a business gets, without agent interaction.
More tangibly, ZCR is a measure of cases or issues resolved using self-service, auto response and Intelligent Process Automation (IPA) described above. These are assisted and augmented by auto triage.
CS leaders can justify the ROI for AI-based automation investments using ZCR. When implemented appropriately these investments can pay for themselves quickly, typically in 3-6 months. ROI can be quantified in terms of dollars by the reduction in the number of agents used, time saved, etc. In addition, there are intangibles that CS leaders cannot ignore like agent morale, customer effort score etc.
Gartner reported that 55% of established companies either have started making investments in the potential of artificial intelligence or are planning to do so by 2020. Also, consumers in the United States are willing to spend 17% more to do business with companies that deliver excellent service, up from 14% in 2014.
In partial automation, Intelligent Process Automation (IPA) can reduce handle time on key processes from tens of minutes to under a minute, reduce human error and most importantly reduce the mundane aspect of processes due to the repetitive nature of tasks. Complete automation which is also known as unattended automation can complete most of this process for a fraction of the cost. A typical process that takes a business between $5-10 per process can be automated under $2 per process using Intelligent Process Automation (IPA).
Calculate your CS Automation Potential with a free ZCR Score Assessment
Get a free assessment of your ZCR Score to understand the universe of cases that can be ZCR’d (resolved without human interaction) in your customer service. Sign up for your free, personalized ZCR Score Assessment today.
With the ZCR score assessment, you can grade your existing processes and assess the potential impact of AI on your CS. The ZCR score is comprised of your self-service score, your automated response score, and your process automation score. Sign up for the assessment now.