Brands looking to tap self-service tools such as virtual assistants for higher efficiencies will need to ensure these systems are powered by the right data intelligence, or they may end up frustrating their customers instead.
When implemented right, virtual assistant tools can be effective in improving customer satisfaction while enabling businesses to better support a team of human service agents who can focus on more critical customer cases. Consumers enjoy faster response time and manpower resources can be freed up to deliver enhanced customer service.
Given the potential, virtual assistants in recent years have seen rapid adoption. Brands were eager to deploy the automated systems and enjoy the much-touted cost savings and operational efficiencies.
These business benefits, however, failed to materialize for many.
According to a report from Forrester, 54% of online consumers believed that interacting with chatbots would have a negative impact on their quality of life. In fact, they anticipated unsatisfactory engagement with chatbots and few had communicated with one that exceeded or met their expectations, Forrester noted.
Gartner's research also estimated that 40% of bot virtual assistant tools introduced in 2018 were abandoned by 2020.
These findings further underscore the need to ensure self-service systems have the capabilities, powered by AI and machine learning, to better address customer cases and offer more effective resolutions.
Why have these applications not lived up to expectations? Chatbots, in particular, often fail to respond to customer queries effectively because they are not connected to the right datasets and systems and are asked to perform rule-based queries. These are very limited, resulting in the inability of chatbots to resolve customer issues.
Furthermore, companies that have not undergone data transformation typically have information that is disparate and stored across multiple locations. This makes it challenging to pull together relevant pieces of data so more comprehensive customer profiles can be established.
In addition, customers will find it frustrating to interact with chatbots that are basic in functionality and trained to solve only a small number of queries. There are only so many times consumers are willing to rephrase their question, just so the chatbot can understand their service issue within its limited vocabulary.
Businesses that are committed to rolling out effective virtual assistants should build such tools with the ability to continuously self-learn, tapping data from various platforms to augment their intelligence.
Connected to different customer systems, such as customer relationship management (CRM), point-of-sales, billing, and social media communications, automated self-service tools will be better equipped with relevant information to address consumer queries.
They can lead to more accurate self-service customer resolution.
AI the necessary brain behind virtual assistants
Organizations that understand this have turned to IBM's Watson Assistant to help them deploy systems that actually deliver on what virtual assistants promise.
Watson Assistant is a conversation AI platform that provides customers with fast, consistent, and accurate answers across any application, device, or channel, including digital and voice.
Using AI and natural language processing, Watson Assistant continuously learns from customer conversations and improves on its ability to resolve issues during the customer's first interaction. It provides consumers with the best customer experience, removing the frustration of long wait times, unhelpful chatbots, and tedious searches.
US insurance service provider, Humana, has reaped the benefits of an AI-powered virtual assistant after working with IBM to roll out a solution that integrated multiple Watson applications in a single conversational assistant. This solution runs on IBM's cloud platform, while Humana operates Watson Assistant for Voice Interaction on-premise.
The insurance provider processes 1 million provider calls every month and its existing interactive voice response (IVR) system was transferring too many calls to human agents. This resulted in a significant cost for Humana and impacted customer satisfaction. Callers were bypassing the IVR systems, even though more than 60% of these calls were routine, pre-service questions that had well-defined answers.
The Voice Assistant was trained using speech customization with seven language models and two acoustic models, each targeting a specific type of user input collected by Humana. The customized training enabled the solution to clock an average 90% to 95% sentence error rate accuracy.
The Watson-based system is able to handle more than 7,000 voice calls from 120 healthcare providers per business day. It can complete an inquiry in about two minutes, without requiring the caller to wait to reach a call center agent.
Powered by IBM Watson, Humana's Voice Assistant has enhanced self-service capabilities that allow the company to more quickly and efficiently provide customers with information they need, across various data points.
Watson Assistant achieves this through machine learning and deep learning capabilities that can accurately respond to user questions, even with relatively small datasets. The IBM solution's AI engine is designed to correctly identify numerous permutations of intent in real-world business and customer interactions.
Watson Assistant has, in fact, helped customers save $23.9 million in benefits, according to a Forrester Consulting study commissioned by IBM. These organizations also clocked a 337% returns on investment over three years, which included savings of $5.50 per contained conversation with Watson Assistant that totaled more than $13 million over three years.
When done correctly, self-service tools such as virtual assistants can prove valuable in helping businesses deliver richer customer engagement and achieve operational efficiencies.
And they are better able to do so with solutions such as IBM's Watson Assistant, which are powered by robust AI and machine learning capabilities to more effectively resolve customer queries.