Customer service has long been a point of friction among the relationships in both B2B and B2C business models. Subpar customer service has continually proven to have adverse consequences on customer retention and revenue, and better customer support is increasingly becoming relevant in the digital age of personalization and catering to customer demands.
New Voice Media reported that businesses lost $62 billion in 2015 from poor customer service alone — a more than 50 percent increase over the previous two years. Additionally, more than 75 percent of customers expect customer service representatives to have visibility into previous interactions — although only half of the agents cite having the necessary context and resources to solve customer issues efficiently.
The problem is that as customers, even in B2B models, expect more personalization, the ability of agents to effectively address those issues within sufficient timeframes is lacking as platforms and software services become more convoluted. Fortunately, AI has developed significantly since its days of limited functionality in customer support chatbots.
From sentiment analysis to embedded AI products for help desks and CRMs, AI is rapidly making an impact on customer support — one of the first markets where we will likely see the most tangible results from AI.
Addressing traditional problems
From a consumer perspective, customer service is one of the most important characteristics of both brand loyalty and purchasing decisions. According to a Microsoft study, 95 percent of consumers indicated that customer service techniques is a crucial factor in their choice of brand loyalty.
In particular, younger generations of consumers are predisposed towards brands with proactive customer service reaches. However, the critical impact of customer service is not solely relegated to B2C models. In many instances, businesses in B2B contexts often overlook the influence that customer experiences have on their revenue generation.
For example, CSO Insights detailed in a comprehensive business sales performance study how nearly 60 percent of enterprises focus on acquiring new customers when they only account for less than 30 percent of the total revenue. The takeaway is that businesses are often missing the picture of just how beneficial improved customer experiences can be on their bottom line.
As SaaS continues its exponential rise, provisioning customer support agents with sufficient context and resources for increasingly complex products has become a significant problem. Human agents are bogged down by repetitive tasks such as answering FAQs or providing logistical information. Further, many brands cite more than half of their customer service requests as falling within the same 6 – 8 topics — clearly revealing the need for more automated efficiency. Enter AI.
Building smarter customer support software
While AI may not be able to replace human agents in explicit cases where social interaction is necessary, such as overly emotional situations, they present some compelling improvements that can help replace repetitive support exchanges and augment support agents on more complex exchanges.
AI customer support solutions are rooted in machine learning (ML) and the ability to efficiently pull relevant data from multiple sources that supplement the job of customer service representatives. From the original AI origins of customer services such as chatbots and online help articles to deflect tickets via self-service, AI-based customer support has come a long way.
Early iterations of AI-based customer support (i.e., chatbots) required significant amounts of maintenance and couldn’t handle complexity outside of a narrow set of instructions. As software solutions became more complex, so did the support tickets — making many chatbots irrelevant or simply an annoyance to customers.
The AI tools can adapt and learn from unstructured data such as historic emails, call logs, past tickets, internal documents, and even how-to videos. The goal is to help route and categorize tickets to the appropriate humans with a relevant data set.
Ongoing development in the field currently centers on integrating these AI products with products that prevail in most business software management systems like CRM, OMS, and inventory management. Smart AI solutions can often answer questions between businesses that involve repetitive inquiries or less convoluted tasks, so one of the primary initiatives has become empowering human agents with better resources. These resources are AI products that are becoming increasingly better at interpreting the context of questions and recommending answers for agents to the appropriate tickets.
AI has reached the point where it is good enough to not only search through centralized data like your online help articles but also through decentralized and unstructured data like past support tickets and conversations. Such response materialization is not simply keyword matching either, AI products can now give customer support agents complete responses — putting answers into the context that they pull from unstructured sources.
For instance, Forethought has built an integrated AI customer support tool, called Agatha Answers, that is embedded into existing workflows like Zendesk and Salesforce. Agatha pulls relevant information from a business’ existing knowledge base, which can consist of sources such as online help articles, internal cloud documents, and information from past tickets. Agatha can understand inquiry meaning based on contextual cues and recommends answers to customer support agents in natural language.
It’s often difficult to imagine how AI products can understand contextual references in languages as complicated as English, but they are becoming surprisingly adept at it. Take the case of OpenAI, the non-profit AI research group backed by Elon Musk: their text generation software was so good at mimicking human writing and predicting the next word in sentences that they actually cited how it was ‘too dangerous’ to release at this stage.
AI’s prominence among customer support solutions and products for businesses is only set to grow, however. According to Gartner, at least 15 percent of all customer service interactions should be handled solely by AI once we hit 2021. Additionally, McKinsey cites how by 2030 roughly 70 percent of companies will have adopted some form of AI, and the majority of companies will be using full-range AI products.
The incorporation of more powerful AI tools that are embedded in modern workflows engenders some vital advantages to businesses. Time to resolution on support issues, even for complex engineering support teams, can be reduced drastically, and as a result, more time and energy is spent on trends that customers expect — such as proactive engagement and personalization.
We are assuredly still in the nascent stages of AI technology, but meaningful development is ongoing in multiple areas — particularly customer service. As consumers and businesses continue to expect more tailored customer support services, evaluating the advantages of AI-based customer support products is a prudent strategy for any forward-thinking enterprise today.