It seems counterintuitive to say that artificial intelligence is a huge boon to customer service. Yet, 37% of major companies use AI in customer service. Isn’t customer service about understanding, listening, empathy, and problem-solving with a personal human touch? Surprisingly, AI is more personal than you think.
Economists, scientists, and IT specialists agree that AI is and will continue to revolutionize the customer service industry through streamlining customer interactions and facilitating substantial savings all to the benefit of happy customers and businesses alike.
Obviously, AI is here to stay. There’s no denying it. Those who don’t embrace it will get left behind. The latest Accenture Technology Vision Report confirms that by 2035, AI will propel economic growth to twice its current figures. Why? Because AI is improving user experiences through improved intelligent interface that is more personalized and natural, downright conversational.
With the world invested in AI to the tune of $126 billion (and rising annually), it’s not hard to see why customer service would be no different. AI promises to improve customer service through bots (“front-end AI-powered bots”) and assistants (“AI-assisted human agents”) that satisfy customer needs, improve employee wellbeing, and save money. The future of businesses that differentiate themselves from the others is in AI.
And with more than 2.7 million American customer service representatives (according to the Bureau of Labor Statistics) earning on average $35,170 per year, it’s no wonder that companies are interested in AI that makes these reps more efficient or, in some cases, obsolete.
In fact, most businesses already do or will use AI in customer service. Front-end bots “converse” with customers directly while cyborgs are human customer reps assisted by AI technology. Both models are common in the industry today. However, six categories of AI assist customer service:
- AI technology or multiple choice chatbots: basic bots with no learning capabilities that converse in simple input-output terms, commonly used by retail and publishing.
- Virtual agents: language-programmed search engines that match keywords to understand inquiries.
- Cyborgs or “human in the loop” bot-assisted agents: assistants capable of efficient conversation under human monitoring to provide customers personalized responses.
- Emotional analysis software in real time: Software that reads customer cues--emotional and social--to respond in real time.
- Intelligent assistants: fully-automated agents that use natural language processing technology.
- Content curation: bots that learn from customer responses or previous purchases and give personalized product recommendations.
Which Services does AI perform?
Where chatbots come in handy is in the repetitive tasks, like providing simple answers to commonly asked questions or password resets. The bots give automated responses to FAQs, freeing up time for the humans to handle more complex queries. In this way, bots alleviate some of the repetition-monotony fatigue humans experience in handling routine tasks, and thus create more specialized human agents.
Time and money saving
Virtual agents working together with humans can adjust to activity spikes that would ordinarily require hiring more staff. So the cost savings associated with hiring and training new hires is significant. On average, it costs roughly $4000 to hire customer service agents, and even more, about $4,800 to train them. That’s why platforms, like industry-specific Watson Virtual Agent, are as great as they require no training with pre-programmed knowledge about the industry upon purchase.
What’s more, automated customer service can work round the clock, on holidays, and during weekends. AI generated customer service is reliable, not subject to human fallibility, like sickness, hangovers, lateness, and anger, when dealing with customers, who get predictable responses to their satisfaction.
For example, China’s WeChat bot fields 1.5 to 2 million China Merchant Bank customer bank balance and charge (mostly) inquiries a day. That’s a lot of manpower reduced to machine! Not to be outdone, Capital One launched the assistant Eno with Natural Language capabilities, and Bank of America’s Erica, will soon assist banking customers.
In the travel arena, the Scandinavian airlines, KLM, uses DigitalGenius that “reads” the customer issue and then suggests a response to human agents to speed up customer response time. The AI learns from the agents’ customized responses taken from the AI preformatted information (macros and clips from past responses), which improves upon the AI initial responses. This allows for a quicker response time (within 24 hours). And Baarb recommends travel spots for customers personalized to customer wants.
But text inquiry-based AI models, like KLM’s, are limited. On the other hand, Natural Language Processing (NLP) in customer service programs like LivePerson (with $200 million in revenues last year), working with IBM’s Watson, allows AI technology to “hear” and route voice calls to the most appropriate recipient. The aim is to alleviate some of the frustration of dealing with bots. The bot handles simple questions; more complicated conversations get passed on to humans and then often relayed back to the bot to finish up. In this way, tasks are delegated to the right technology or human, creating a more efficient process--including the ability for reps to multi-chat.
AI and CRM
Given AI’s expansion to 38% of all business applications--and expected to double next year--AI for Customer Relationship Management is a natural outflow. So much so that Salesforce (featuring AI-powered Einstein CRM) estimates AI for CRM will increase business revenues globally by $1.1 trillion come 2021.
Salesforce uses Einstein Supervisor to monitor call centers, tracking patterns in calls and data regarding product purchases. Also, basic ordering and information are fully automated on this platform, tasks that would not crush an organization with AI mistakes, should there be any. In fact, popular food chains like Subway and Starbucks use AI for placing orders. Here, AI potentially replaces hostess and salespeople tasks with bots and computers.
Predictive insights and analytics
Predictive insights and analytics that enable agents with better insights into customer behavior to anticipate customer needs and improve satisfaction lie in programs like Zendesk’s Satisfaction Prediction. Also, machine learning helps narrow down best responses for particular issues, providing agents with new insights and novel approaches to problem solving.
A CRM platform with AI capability optimizes user experience by analytics and predictive insights that is critical to customer retention. It allows customers to easily interact with a company, make purchases, and get help if they need--with time-saving results.
CRMs with AI can also conduct pre-emptive customer service, detecting customer potential problems before they arise by scanning websites and in-app activity and respond in real time with virtual service agents. This prevents customer frustration and purchase abandonment rates
Finally, machine learning can provide unparalleled personalization that ranges from product suggestions based on buying habits to websites tailor-designed in real time to meet customer reading and browsing habits--all resulting in greater customer experience and thereby greater conversion and repeat business rates.
Customer service isn’t the only area where Artificial Intelligence is playing more and more of a role. Virtual personal assistants like Cortana and Siri use AI to help search for the information you seek and AI is playing an increasing role in SEO, with Google using an SEO algorithm called RankBrain to mimic human preferences and help determine where websites are ranked in the search results – partly based on your previous search history, as well as factors that are specific to each website’s SEO.
You may also use AI every day without even thinking about it – with smart home devices, in video games that generate increasingly life like characters and it won’t be long before self-driving cars become standard.
So, will AI replace humans?
No, not right now, at least. Though most executives believe technology will reduce the customer service force by 4%, not all customer requests can be handled by a robot. And though AI-powered conversational interfaces are revolutionary innovations for the customer service field, there are still interactions, like financial questions, that are too complex for AI.
For example, AI still falls short on voice-based communications. Interference, such as background noise, accents, mumbling, and unusual speech, create AI translation errors.
However, the startup Cogito created a tool capable of reading customer emotions, by reading content and tone. When tested on insurance provider, Humana, Cogito claims their system improved customer service and reduced call backs by 10%.
And yet, the more AI gives, the more people want--more. Interestingly, Microsoft’s 2016 State of Global Customer Service Report proves demand and expectations for technologically-enhanced customer service creates higher expectations for customer service agents to do more, especially by millennials who expect customer service agents to know their contact and product information before contacting them. Thus, higher expectations produce demand for more personalized and specialized service, and thus, a higher niche of customer service.
And then, there’s the increased IT needs that come with improved customer service and AI integration, an entire industry that likewise must improve services to meet the demands of personalized customer service. Companies like London-based Prosyn, an IT support company in London that concentrate on personal services to power AI-providing tech companies up the personalization ante in customer service. In other words, companies that service AI-powered technologies also thrive with the new AI customer service model.
Of course, the catch-22 is that the more capable AI bots become at their jobs, the more likely they’ll take over human jobs.
While speculation about job loss, force reduction, or reallocation of resources into more specialized support abounds, customer service presently stands to gain from AI. The ultimate benefit to customer support human agents, however, well, that’s a double-edged AI sword.