As customers and companies become more familiar and comfortable with artificial intelligence, the conversations around it have become less “pie in the sky” and more along the lines of “oh, that’s how it can be used”. The speculations into how we might work alongside our robot coworkers and the potential of machine learning won’t end anytime soon, but there’s a lot to be said about practical applications of AI technology. Especially in relation to the customer experience.
This is great news for businesses, since it’s much easier to tie realistic use cases for artificial intelligence to their goals. But experts advise that companies should be careful not to implement AI for AI’s sake – they argue that artificial intelligence will be most successful when it facilitates humans to deliver better customer experiences.
That’s why many analysts are encouraging the adoption of AI tools with a calculated approach, one that focuses on “how” and “why” the tools provide a better overall customer experience. Forrester, for example, advises that “having a successful AI-driven customer service or sales program will depend on the processes that support a blended AI approach. Humans will play a critical role in the ongoing optimization of AI.”
For those looking to be ahead of the curve in implementing AI for better customer experiences, here are five ways it can be leveraged right now:
Customers who would rather help themselves instead of interacting with support agents may sometimes find themselves digging around for a solution. The self-service process is borderline tedious when it entails a combination of a Google search, navigating an online help center for a relevant article, and then confirming that the article’s solution adequately addresses your issue. Some customers, especially the non-tech-savvy, may find that kind of self-service arduous and unfitting for a smooth CX.
Recent innovations in artificial intelligence can both alleviate the hassle of customers searching for help articles and ensure that they’re given the right information for to solve their problem. AI utilizing machine learning and natural-language processing (NLP) is capable of learning which help articles can best solve a customer’s problem and recommending the appropriate article to the customer. Not only that, but customer experience leaders can determine where it makes the most sense for customers to encounter this kind of automated self-service; be it in front of a help center, at a critical point in the buyer’s journey, or on a mobile website or app.
Speaking of help articles: as products and services become increasingly complex, support organizations are finding it difficult to keep their help articles relevant and up-to-date. Few things are as frustrating to a customer as unhelpful support content, and proactively preventing that means having an actionable content optimization strategy in place. If customers are bouncing quickly from a help center or the articles within it (or they say how an article failed to help them in a follow-up), that’s a likely indication that the content wasn’t adequately tailored to their need or issue.
Luckily, there’s AI that can support the creation and actualization of better tailored content for a specific customer base. Deep learning models can catch the common words and phrases related to specific issues found in support tickets, and then can make tactful recommendations for optimizing help center content. For example: if customers are submitting support tickets with the subject “change my password”, the AI will recommend appropriate editorial adjustments to the related help article entitled “How to update your login credentials”. The article can be changed to reflect how customers communicate their issues and make the content easier for them to find and understand. By providing a support organization with insight into the issues their customers are facing and how they communicate them, content creators can communicate solutions that align with a better customer experience.
How many times have you heard a support agent say “Let me check on that for you” because they simply don’t know the answer to your question? Support agents generally spend 20% of their time on the hunt for product information, which draws out a support interaction and may have a negative effect on customer satisfaction.
The same artificial intelligence that automates self-service suggestions to customers can be utilized by agents, too. Let’s say a customer is locked out of their account from too many failed password attempts, but they desperately need access and submit an urgent support ticket. The agent who received the ticket isn’t familiar with the internal process for unlocking an account, so they need review the appropriate internal documentation… but first, they need to find it. The right AI tool can analyze the support ticket and recommend a relevant article from the company’s knowledge base, and can do so directly in the agent’s interface. By enabling agents with the right information when they need it, they can efficiently address their customers’ issues soon after they arise.
Our digital activities and interactions result in tons of data for machine learning algorithms to utilize; it’s essentially the fuel for AI’s predictive proficiency. Ever wondered why machines are so good at answering questions like “what’s the fastest route for my commute home at 6:00 PM on a Thursday?” By aggregating numerous trips taken by those who’ve traveled a similar route, the AI can determine a well-researched, real-time recommendation for getting you home quickly.
Similarly, the data recorded from customer service interactions can be utilized to improve CX. By assessing the details of previous support tickets, an AI tool can understand whether a current support interaction will lead to a positive or negative customer experience, resulting in an accurate customer satisfaction (CSAT) score prediction. Those details that can impact CSAT scores might include the length between the first reply and subsequent response times, how much effort is put into resolving the customer’s issue, and if text responses with similar wording have resulted in satisfied customers. This is the kind of AI application that doesn’t replace agents, but instead augments their efforts to deliver better customer experiences.
Of course, one of the most compelling value propositions of automation is that it frees up time for humans to focus on other high impact efforts. Artificial intelligence is already enabling businesses to improve their customers’ experiences in ways they haven’t been able to before.
Dollar Shave Club, the on-demand shaving razor service, benefits from the extra time that Zendesk’s Answer Bot affords them to optimize their customers’ experiences. With the time they’ve saved through automatic ticket resolutions, they’ve been able to:
- Create a “Help Center Task Force” that ensures their help articles are constantly relevant and up-to-date for their self-serving customers (and to supplement Answer Bot’s recommendations).
- Increase the amount of time that they offer live chat support throughout the day, permitting them to offer more real-time assistance without increasing headcount.
- Launch an internal monthly e-newsletter on customer engagement insights that highlights trends and keeps their agents better informed on their success metrics.
- Find bandwidth to launch a “Test & Learn Team” to try out new email messages that might improve their members’ customer experience.