Why Conversation Orchestration is essential to great AI driven customer service
Since the beginning of this year, barely a day has passed without AI and ChatGPT grabbing headline news. These technologies are profoundly and permanently impacting our society. In this blog, we’re eager to share our insights on how the rapid adoption of AI and ChatGPT is changing the traditional Call Center and CCaaS (Contact Center as a Service) industry. We will elaborate on the 5 reasons why conversation orchestration is essential by facilitating an effective Human-In-The-Loop (HITL) as the last mile to make the customer experience seamless and frictionless.
1. AI and ChatGPT are powering the next big leap in customer self-service.
Conversational AI platforms are catalyzing a significant transformation in customer self-service. These technologies offer vast enhancements over the previous generation of chatbots, which previously frustrated customers with misunderstandings and irrelevant responses. The new generation of chatbots, powered by AI and ChatGPT, as a subset of Large Language Models (LLMs), excels in generating human-like responses that effectively address a broad spectrum of customer queries. These advanced chatbots can provide tailored responses to meet specific customer needs, leading to shorter wait times and enhanced overall satisfaction. They possess the capability to learn autonomously from past conversations, equipping them with extensive knowledge from data sources. As such, it is no longer a case of if CX organizations will be forced to further invest in conversational AI, but when. Ensuring that this new wave of technology is properly architected to deliver for customers is essential.
2. Digital self-solutions are not yet delivering their full promise.
Customer Service executives are confronting a perfect storm of challenges, including increased call volumes and a talent shortage. Add to this mix inflated customer expectations and high-effort service designs, and it becomes easy to understand why they are under significant pressure.
Digital and Customer Self-service solutions have emerged as an increasingly prevalent tool to address customer service challenges and to optimize operational efficiency. CX Leaders often perceive these technologies as a win-win, offering cost savings and delivering a great experience through quick resolutions. There is research to back this – with Salesforce’s 2023 state of Customer Service suggesting that 59% of customers even prefer self-service tools for simple questions/issues.
So, what’s the problem you may ask? With 88% of customer journeys starting in self-service touching multiple channels according to the latest survey of Gartner in December 2022, only a shocking 13% of customers are fully able to resolve their issues through self-service as the only channel.
Gartner
The same survey of 1,492 B2B and B2C customers found that a seamless transition between channels is a key driver of positive customer experience (CX) outcomes and future self-service adoption. When the transition is high effort, meaning that switching channels did not lead to resolution, the probability of positive CX and loyalty outcomes is dismal. Conversely, 74% of customers who experience easy transitions to a customer service representative say they’ll return to self-service next time. Ensuring an easy transition between channels, and particularly to a live agent, is critical to deliver on the value of self-service/AI for both customers and businesses.
“Despite resolving their issue in assisted service, these customers are just as likely to use self-service in the future as those whose issues were fully resolved using self-service. Eric Keller, Sr. Director Analyst, Gartner”
3. Effortless Transitions Improve Business KPI’s and Customer Experience.
The above survey highlights the benefits of promoting seamless transitions between self-service and assisted channels. Seamless transitions improve Customer Experience across a range of metrics and boost loyalty, but they also save time and resources. In fact, 93% of respondents reported high customer satisfaction (CSAT) when there was a seamless channel transition.
4. Human-In-The-Loop (HITL): Desired or Required When It Matters
When designing customer self-service, we should treat the Human-In-The-Loop (HITL) as the last-mile to make the customer service experience seamless and frictionless.
In short, there are several occasions where we need to bring in the knowledge, flexibility and skills that only a human agent or knowledge worker can deliver:
a) Complicated issues where self-service is not able to drive the issue to resolution.
When customers need to modify their contracts, they often require the expertise of a live agent knowledgeable in contract management and legal matters. Instances where a bot might ask too many clarifying questions, leading to customer annoyance, further highlight the need for human intervention. Additionally, complex issues, particularly in banking or other specialized areas, require support of knowledgeable professionals for the foreseeable future.
b) When detecting a buying or a defecting customer signal
Companies keen on not missing revenue opportunities must be adept at identifying the right moments for a live agent to step in—especially when a customer shows interest in making a purchase or exhibits signs of potential departure. In such instances, it’s crucial for the bot to recognize these signals promptly and insert the most relevant salesperson or expert with the highest priority through real-time triaging.
c) Human Empathy
Early detection of customer sentiment allows businesses to capture their customers’ feelings and respond appropriately via live agents. There’s a common belief that bots lack empathy and that demonstrating empathy effectively requires promptly connecting customers to a human agent. However, some staunch technologists argue that achieving empathetic responses through technology is only a matter of time. Our research indicates that customers tend to become more irritated with perceived ‘fake empathy’ from a bot than when interacting with a human being. Therefore, ensuring a seamless transition to a live agent when it truly matters is crucial.”
Take, for example, an insurance company in the Netherlands that has been utilizing conversational AI with success. Despite customer satisfaction with the bot, the company occasionally routes customers to live agents as a gesture of care. This inclusion of a human touchpoint, even briefly, reinforces customer care before handing back the conversation to the bot.
So, Bots aren’t good at delivering empathy to customers, but they’re increasingly good at understanding how the customer feels. You don’t want the AI to deliver the empathy, you want to leverage the AI to both prime the agent to know the ideal tone to set, and to pick the agent best able to deliver that as well.
In an era of advanced AI and ChatGPT, customers expect immediate and accurate responses. Yet, many existing technologies only provide basic First-In-First-Out or are leveraging multiple queues in traditional call centers as hand-offs from bots to human agents without any ability to manage traffic, prioritize crucial conversations, or seamlessly connect users with specialized knowledge workers. This erratic transfer process often results in the frustration of unpredictable waiting times and in some cases agents who are unable to resolve issues within the first contact. As a result, the existing disconnect between bots and human agents is becoming increasingly undesirable in today’s customer-centric landscape. Providing the opportunity for a simple escape from a bot to a live agent is essential. This fact has led to the necessity of what we now call conversation orchestration. In an ideal world, perfect harmony should exist among digital self-service channels and live agents or knowledge workers for an optimized customer experience.
5. Conversation Orchestration: the missing piece of the puzzle in AI enhance Customer Experience?
In addition to its application in conversational AI, Generative AI also plays a significant role in AI Agent Assist, a technology that supports human agents during customer interactions. By providing real-time suggestions and relevant information, ChatGPT can help agents respond more effectively and efficiently to customer inquiries. However, what Conversational AI and AI enabled Agent Assist are not resolving is the absence of harmony between automated responses and live agent interaction.
Here at SentioCX, we envision a future where transitioning conversations from digital self-service solutions to live agents is effortless and seamless. The future of customer service will involve chatbots and live agents working together in harmony to provide the best outcome to customers and agents. Conversational Orchestration leverages AI to facilitate agent-pairing and to orchestrate conversations. This enables organizations to take advantage of conversational AI and operational efficiency while also optimizing customer intimacy. In Ronald’s previous blog “Resolving the Customer Experience paradox”, the point was made that current routing algorithms in CCaaS solutions are outdated. They cannot make routing decisions based on recognition of real time intents or sentiment from the bot. Hence, we came up with a new way of revolutionizing Human-In-The-Loop. We use AI based agent pairing to create seamless and effortless transitions from digital self-service to live agents and knowledge workers.
The essential components of AI enhanced Customer Experience
By using the three pillars of AI enhanced Customer Experience combined with AI and ChatGPT, we provide an evolution on the now obsolete traditional IVR. Users can effortlessly ask questions and receive automated responses or when necessary be connected directly to the right agent or knowledge worker to solve their issue.
In summary
Conversation Orchestration effectively prioritizes high-value interactions, through predictive actions based on agent availability, skills, and proficiency. Furthermore, it dynamically escalates priority with changing conditions or sentiment and temporarily provides self-service or follow-up for lower value/priority interactions when service levels are exceeded. This approach is designed to optimize the impact on key performance indicators (KPIs) such as customer churn, customer satisfaction, and (up and cross) sales. SentioCX delivers these capabilities out of the box along with dashboards to manage, track and visualize the impact on the most relevant KPIs.
Sentio Insights provides dashboards and reports to manage, track and visualize KPIs and Service Levels
Resolving the Customer Experience paradox
Ronald Rubens
My experience in the world of Customer Experience over the last 25 years
by: Ronald Rubens
Today are both exciting and alarming times in the world of customer experience. Over the last 25 years, I have seen how new technologies can create extra-ordinary experiences for customers and organizations. However, the reality on the ground is often different as organizations in their quest for optimizing operational efficiency, have applied a variety of technologies, such as bots, to take over where skilled people and knowledge workers once excelled. These bots turned out to be effective by successfully managing customer interactions, but also created friction or revenue loss when customers expected immediate access to a live agent. The good news is that I believe this old-age paradox of customer excellence versus operational efficiency is finally behind us; if organizations are doing the right things, they can now create extra-ordinary experiences while benefiting from operational efficiencies. However, before we get to the answer on how to achieve this, let’s have a look at some current facts and allow me to take you back in history.
It was only the other day that I had to wait for more than 24 hours to get an answer from a major airline on an urgent WhatsApp request, impacting an important customer visit. The next day I had to contact my energy company to renew a contract, but could not find the energy company’s phone number on their website. Given the recent developments on energy, the energy company took the decision to hide their customer service phone numbers with the hope that most interactions could now be taken over by a bot. And speaking about bots, who has not recently interacted with a bot to find out that the bot didn’t understand our question or was not able to transfer us to an agent? Or when the bot finally agreed to transfer us to an agent, we had to repeat the information all over again or found out that the customer service agent was not the right person to help us out.
It seems that I was not alone in these challenges. Hold-times in call center queues have increased by more than 50% over the last two years. Despite attempts to reduce hold-times by using self-service capabilities, such as ‘interactive voice response’ and bots, 59% of all consumers feel companies have lost touch with the human element of customer experience. In a recent survey by Zendesk, 54% of respondents say it takes too many questions for the bot to recognize it can’t answer the issue and then they must start all over with a human agent. As a result of these recent experiences 61% of customers would now defect to a competitor after just one bad experience as expectations around customer experience have risen in the digital age. So, what can we learn from history?
At the turn of the century, I worked in the United States with a talented group of Bell Labs Engineers in the contact center division of what is now called Avaya. My manager at the time gave me a book titled: “The Discipline of Market Leaders” by Treacy and Wiersema. It was one of these New York Time best sellers, which instantly became a success as it described three paths to market leadership. These three paths – also known as value disciplines – were customer intimacy, operational excellence, and product leadership. In essence, a company had to pick one of these disciplines while meeting industry standards in the other two. To become ‘Masters of two’ of these disciplines was a real challenge as inherent tensions would appear between the operating models that each value discipline demanded. In the contact center and customer experience space, this led to a so-called customer experience paradox as customer intimacy and operational efficiency were hard to unite under a single all-encompassing value discipline. Companies had to deal with trade-offs between the quality of the customer experience and the costs of providing it. This trade-off always seemed baffling to me, and a life-long quest started to explore how we could break this paradox and to bring both customer intimacy and operational excellence in balance.
Twenty years later, encouraged and inspired with the famous words of Steve Jobs: “You’ve got to start with the customer experience and work your way backwards to the technology”, I started a discovery of the fundamental issues in the customer experience space and founded what is now SentioCX. “Sentio” is a Latin word for “I experience” or “I feel” to emphasize the starting point of customer experience. Our vision was to find a way to have technology – in this case bots – and humans work together seamlessly, to create extra-ordinary and frictionless experiences for the customer and as a result create an optimum between operational efficiency and customer intimacy. Today, we believe that we are finally able to unite these two paradoxical disciplines.
Our vision at SentioCX was to have companies, who were using bots and conversational AI technology, benefit from cost savings while at the same time improving customer retention and satisfaction by providing a seamless transfer from a bot to a live agent. It seemed simple and straightforward as I write this, but the challenges to make this happen were multiple. How do we make it seamless for customers? How can we ensure that questions from customers get solved the first-time? When is the right moment for the bot to hand-off the call to a live agent – in some cases pro-actively? How do we make sure the most qualified agent gets a specific call? The proliferation of Artificial Intelligence and a variety of omni channel concepts brought many powerful use-cases for Contact-Center-as-a-Service and Conversational AI technology, except providing a seamless transfer from a bot to a live agent. At best, agent hand-offs from a bot were ‘blind transfers’ by throwing a ball over the fence or launching chats in agent queues with the hope that the agents would pick out the most relevant chats. So, what was the answer?
Along with deep integration with bots, conversational AI applications, service management and live chat applications, we discovered very early on that we needed ‘a routing algorithm’. After doing extensive research of ACD (Automatic Call Distribution) routing algorithms known today in the contact center and CCaaS (Contact-Center-as-a-Service) space, we discovered that most routing algorithms were too basic in the sense that they could only route chats or voice calls based on historical importance of the customer (retrieved from a CRM system) or IVR (Interactive Voice Response) choices. Hence, we couldn’t use these existing algorithms to facilitate a seamless hand-off to an agent.
The customer experience we wanted to provide was that by applying Artificial Intelligence, the bot would understand the customer and more importantly, detects ‘early on’ that it doesn’t understand the customer or believes that human intervention is required. It would then pro-actively with priority and/or urgency seamlessly hand-off the chat to the right agent with the right skills and proficiency levels or if necessary to a knowledge worker with a specialized skill.
What we wanted to accomplish was apart from the customer profile, to classify and prioritize real time intents captured by the bot and based on these intents match the chat with an agent or knowledge worker with the relevant skills and proficiency levels and dynamically adjust the routing priority. So, what we did was to start inventing a ‘brand new approach’ of a sophisticated ACD (Automatic Call Distribution), which we now refer to as our ‘Intent-based Intelligent Routing engine’. And so it began: we filed a patent, simulated the algorithm, validated it, and started to build the product by hiring a bunch of talented software engineers who were as passionate about this business as we were.
During the time of Tracey and Wiersema’s publication of their book, customer experience managers had a challenging job: they had to choose between operational efficiency and customer experience. Today we believe that with the advent of technology, and our work at SentioCX, we are now able to unite these two paradoxical disciplines. I am excited about the future. Not only for our company and the difference we are making, but also for the future of the CX industry and the end-customer. I believe that with the right actions, we can create a perfect symbiose of technology, bots, call center agents and knowledge workers, working to create great experiences for customers and organizations.