Can Chatbots and AI Reduce Customer Support Costs?
Mike Mangi | Director of Digital Strategy
September 28, 2017
Major brands like Staples, Dominos, and, Capital One, are investing in chatbots to provide customer service support. Chatbots are able to provide 24x7 real-time, cross-channel access to a virtual customer service agent. And these brands believe that chatbots are a way to increase customer service scores, build brand loyalty, and increase sales. And you may be asking yourself if they can do the same for you.
In order to answer this question, let’s take a look at what is behind chatbot technology, how some leading brands are leveraging this technology, some common pitfalls to avoid, and what you can do to begin developing a strategy that uses chatbots to meet your customer support needs and potentially lower support costs.
What exactly are Chatbots?
- Chatbots are computer programs designed to simulate human interaction. Some of them operate via text message inside platforms like Facebook Messenger, Slack, and Kik, and others like Siri, Alexa, and Cortana, are able to understand speech through natural language processing.
- They can be programmed to respond to a limited set of commands or keywords, or they can be extended with artificial intelligence and machine learning, to continuously expand and optimize their response set.
- Business Insider predicts that chatbots have the potential to save businesses billions of dollars per year in customer service costs.
Can Chatbots Solve Customer Service Problems?
Customer expectations are high. One report found that 42% of consumers who reach out to brands online expect a response with an hour. While a study by Sprout Social found that retailers are only responding to 19% of consumer questions, and that the average response time is over 10 hours. And you thought hold times on phone support were long! This illustrates a huge gap between customer expectations and brand responsiveness.
A recent study by Aspect Software, found that 71% of consumers want the ability to solve most customer service issues on their own. And 61% think that charbots allow for faster resolution or answers.
A poll Retale published this year found that 70% of 500 millennials who interacted with chatbots had a positive experience.
And a recent Oracle survey found that 80% of sales and marketing leaders say they already use chatbots or plan to do so by 2020. That means we have hit a tipping point where brand chatbots will be a baseline expectation for customers.
How Can Chatbots Help You Achieve Address Your Customer’s Needs?
Chatbots are able to quickly respond to common requests, like product availability, sizing information, and average shipping times. They can provide information about store locations and hours, and register customers for email or text message alerts to notify them when there are coupons or sales.
Many chatbots currently limit the way customers interact with them. Rather than requiring customers to learn a set commands, the chatbots use links and buttons as guides.
Sephora’s chatbot allows customers to book a makeover, share store feedback, or contact customer service.
Dom, the chatbot from Dominios can place and track orders.
Shake Shack’s new chatbot can answer 300 customer questions, but is not yet ready to take orders.
HubSpot is using a chatbot to explain the recent acquisition of chatbot builder Motion AI.
Third-party chatbots are popping up to help individuals who rent or sell products and services online. Hostbot provides a way for Airbnb and VRBO hosts to answer guests questions like the WiFi password, and how to access the laundry room, at any time, day or night, without having to answer the phone.
Expanding Versatility of Chatbots
Using APIs to connect chatbots to other data sources can expand their utility. Retail chatbots can connect to shipment tracking data to tell you when your purchase will arrive. Travel bots can connect to weather data to help you pack. Connections to social media platforms make it easy to ask a bot to share for you.
Chatbots can provide personalization by collecting preference data like sizes, favorite styles, and important dates like birthdays and anniversaries to send reminders. Capabilities will continue to grow as they integrate with Artificial Intelligence (AI) and Machine Learning (ML) engines and data about frequently asked questions can be analyzed to better anticipate customer needs.
According to McKinsey, 35% of Amazon purchases and 75% of what Netflix customers watch comes from recommendations based on predictive algorithms that correlate past customer purchases, searched product, and what others have purchased to determine what should be shown to any given consumer.
AI and Machine Learning with sentiment analysis can be used to detect mood, tone and frustration level, and perhaps trigger human support intervention.
Pitfalls to Avoid
Of course quality counts. As the novelty of chatbots gives way to repeat use and the greater need for utility, brands have to be careful not make the process of interacting with their chatbots cumbersome.
There can be a high cost for not getting this right. According to a recent survey from the International Council of Shopping Centers, 82% of consumers will look elsewhere if they experience poor customer service.
This past Spring, Facebook admitted many of its bots were only understanding 30% of requests. Causing some brands to remove their bots from Messenger.
You may remember the backlash against Microsoft’s Clippy. Which should serve as a cautionary tale to make sure your chatbot is useful and non-intrusive.
Remember that chatbots are an engagement with your brand. Carefully consider the voice and tone of your chatbot. It is a brand representative and needs to feel like it’s representing your brand. The friendly tone of the eBay ShopBot may not be appropriate for a financial or news organization.
Chatbots and Tier 2 Support
A report by Accenture outlines benefits beyond customer service cost reduction noting that increasing first response time and customer convenience leads to increased customer satisfaction scores.The report references a customer service chatbot pilot test that was able to resolve 82% of interactions by itself and 88% when combined with a human customer service agent.
Human agents will continue to be an important part of holistic customer service support. Chatbots can be thought of as augmented support. When a chatbot encounters an issues it does not know how to resolve, it can pass the customer on to Tier 2 human support.
Aspect software found that human customer service agents prefer to handle more complex support issues.
Capgemini reported on a recent success metric of 35% improvement in customer service efficiency for Dutch airline KLM when they started using AI to assist their human agents.
Chatbots can reduce customer support costs, and provide a better customer service experience, as long as they are optimized for usability and paired with human support for issue escalation.
The chatbot ecosystem has become quite mature. There are many tools and services available for creating chatbots. Amazon has Lex and IBM’s Watson works with Botkit. Google recently launched an analytics service just for chatbots called Chatbase which will help developers optimize their bots’ user experience.
Mastercard is providing an API to integrate their Masterpass Chatbot into popular messaging platforms like Facebook Messenger, Slack, and Skype. This will allow developers to more easily integrate payment processing into their chatbots.
Ultimately, integrating chatbots into your customer support strategy will enhance the user experience for your customers, free up your support team to focus on important issues, and save your business support costs.
If you’d like to read more about enhancing your customer support site, check out these content management tips that will improve user experience.