Scale Customer Support with Conversational AI Agents
The COVID crisis has left many customer service leaders faced with a clear challenge: scaling their support organizations without adding headcount or risk losing their loyal customers. This has led to a fundamental re-thinking of the self-service strategy that relied on technologies such as search, chatbots, and knowledge portals to make it work.
The problem with these traditional self-service channels is that it is focused on deflection rather than resolution. There is a subtle but powerful difference between both of these viewpoints. Deflection is focused on vanity metrics like reducing your call volume into your contact center and resolution is focused on north star metrics such as auto-resolving your issues while providing a great experience. Let’s look a little closer at why technologies like search, chatbots, and knowledge portals fell short.
Legacy Chatbot and Manual Self-Help Knowledge Base Search
Legacy self-help search experience expects customers to sift through pages & pages of knowledge articles to get to the answer they were looking for. At a time when you get recommendations from everything from what food you want to order to what movie to watch, why are consumers expected to search for a needle in the haystack of a company’s knowledge base? Moreover, there is no ability to dialog with search results, they are not personalized or sorted based on the user’s profile leading to poor adoption.
Chatbots, on the other hand, were deployed with high expectations but have failed to deliver on the digital promise. The biggest challenges – they are highly scripted if-then-else flows that fail the moment users go off-script. Secondly, they are super expensive to build and deploy. Just look at Erica, Bank of America’s virtual assistant. At an estimated cost of $30 million that took 100 people. Chatbots are only as good as your knowledge base and more often than not knowledge bases are not structured or organized. Oftentimes each department ends up authoring their own knowledge articles results in inconsistencies or sometimes including repetitive information making it difficult for your customers to find the resolution so they end up abandoning and switching to a live channel.
Why Conversational AI Agents Are Important
When customers do end up switching channels from a self-service to a digital channel, critical information from the customer journey such as what knowledge article did the user read before escalating, when did they last login, from what device, location, etc. is lost and is not available for the service agent at the time than engage with the customer so they end up repeating the resolution options leaving customers highly unsatisfied with the overall experience.
A new era of Conversational AI solutions that promise to deliver on instant resolution of customer support issues (not just deflection) with a highly engaging human-like dialog experience is emerging. These solutions not only come built-in with customer domain-specific ontology – for example, there could be 1000 different ways customers can request how to auto-renew their subscription but also use AI to auto-triage and recommend resolutions for agents thereby drastically improving agent productivity.
How to Get Started with Conversational AI
When considering the implementation of Conversational AI into customer service, questions regarding knowledge-base transfer and time-to-deployment raises many concerns for organizations. How long will it take for Conversational AI solutions to obtain the vast amounts of data from manuals, FAQs, guidelines, and other knowledge bases? Will knowledge bases have to be manually updated every few weeks?
Fortunately, with Conversational AI solutions, service agents skip the hassle of authoring knowledge articles or manually transferring data from one system to another. Instead, Conversational AI solutions leverage unsupervised Cognitive AI to seamlessly integrate with existing knowledge repositories. By quickly ingesting historical data and plugging right into existing enterprise systems, Conversational AI solutions eliminate the need for training, prep-work, and data cleansing, allowing issue resolutions to take effect almost instantaneously.
In addition to smooth and timely implementation, Conversational AI solutions continuously learn by processing data and accumulating knowledge from every interaction, eliminating the need for constant updates or human intervention. The self-learning capability of Conversational AI solutions allows service agents to focus less on managing systematic issues and more on meaningful tasks that bring value to the business.
When service requests do not match any existing information, Conversational AI solutions automatically identify and close these knowledge gaps by searching through external knowledge bases. For each new piece of information obtained from external sources, the internal knowledge base is updated and enriched accordingly, keeping Conversational AI solutions up to date with current and fresh data to perform tasks and resolutions successfully.
Deploying Conversational AI Agents
Conversational AI solutions benefit not only employees and enterprises but customers as well. Leveraging unsupervised natural language processing (NLP) and natural language understanding (NLU) capabilities, Conversational AI agents recognize and understand the intent behind a user’s query regardless of how the question is worded, allowing them to generate immediate and relevant responses to customers. When searching for answers or recommendations, customers no longer have to worry about using the right keywords, sifting through endless channels, or being left high and dry with their issues unresolved. Instead, customers have the freedom to interact naturally with Conversational AI agents and trust that they will be provided with accurate responses, making their experiences smooth, personalized, and worry-free. These experiences will subsequently drive higher customer satisfaction levels, empowering them to return for future interactions.
When switching from channel to channel, user information from previous conversations, articles, and devices carry over to current interactions. By remembering user information, Conversational AI agents provide smooth and consistent experiences across all channels from start to finish, leaving zero room for delays or disruptions. This, in turn, helps customers avoid having to repeat their questions or search through old requests.
The ability to self-learn in real-time is also an advantageous feat of Conversational AI agents. By continuously learning from past interactions and ensuring knowledge bases are updated, the prediction accuracy of Conversational AI agents improves, expediting future issue resolutions for users. As wait times become more reduced and responses more accurate, customer trust and loyalty are built and maintained, making your brand one to follow amongst competitors.