Customer Service Automation: Tips to Succeed and Mistakes to Avoid
When done right, customer service automation creates a better experience for customers and agents alike.
The phrase “customer service automation” evokes different feelings depending on who you are. As a consumer, you may have a love/hate relationship with automated customer service: loving it when you want a quick answer without asking a human, and hating it when it doesn’t work. Service or CX leaders may view call center automation as a necessary evil to handle growing contact volumes at scale. Call center agents may view support automation as a threat to their job.
Love it or hate it, automation in customer service is here to stay.
So let’s take a closer look at what support automation really means, why it’s actually a good thing, some best practices to embrace, and some mistakes to avoid.
- Customer Service Automation: Our Present and Our Future
- Automated Customer Service Definition
- Key Drivers Behind the Support Automation Trend
- Examples of Call Center Automation Technology
- 7 Benefits of Customer Service Automation
- 7 Best Practices for Implementing Support Automation
- 5 Common Mistakes with Customer Service Automation
- How Astute’s Smart Automation Improves Efficiency and CX
Customer service automation gets a bad rap, but is it entirely deserved? When applied well, call center automation technology can do amazing things, improving outcomes for customers, business leaders, and agents alike. And these days, embracing automation is the only way to stay competitive.
As Forbes Contributor Brad Birnbaum explains, support automation “helps to streamline the ever-increasing volume of interactions across touchpoints.” That’s why the use of AI in customer service is expected to increase 143% in 2020. But Birnbaum also cautions: “More [automation] isn’t always better, especially if it compromises the overall customer experience.”
With automation, more isn’t always better.
So how should you strike a balance between these seemingly conflicting objectives? It might help to start by better defining the idea of automated customer service.
Automated customer service, sometimes also called call center automation or contact center automation, is customer support that relies on technology to complete tasks instead of human agents. Entire processes or just parts of processes can be automated. Call center automation technology often includes a combination of narrow AI, robotic process automation (RPA), machine learning (ML), and natural language processing (NLP).
Contrary to popular belief, automating a task does not mean a human is put out of a job. In fact, automation is an important tool in the forward evolution of the world’s workforce because it frees humans from the repetitive and the mundane, freeing them to apply their uniquely human talents to more nuanced and complex issues. At Astute, we think about call center automation technology as automating the process, not the people.
The need to automate customer service processes is driven by consumer preferences and expectations. In a digital age, customers expect immediate 24/7 help with issues and questions, yet also expect those experiences to be highly specific and personalized. This is a key factor behind the rise of live web chat and mobile messaging as the preferred methods to communicate with companies. Recent research from Twilio found that 89% of customers prefer using messaging to communicate with companies – but surprisingly, only half of companies offer service on this channel.
Nine out of ten customers prefer messaging to contact businesses, but only half of companies offer this channel.
Customer service chatbots were created to address this challenge. And their popularity is growing: in a 2020 Astute survey, bots were in the top three choices of how consumers prefer to engage with customer service.
Automation is a broad concept with many specific applications in the world of call centers and customer service. Here are a few common applications of contact center automation tools:
- Chatbots for customer self-service across live chat, SMS, and social messaging
- CRM process automation/agent assistance, such as automatic case creation, next best actions, and auto-suggested knowledge content
- Inbound email processing and automated responses
- Social media monitoring and social case routing
- Conversational IVR
- Customer portals and knowledge bases
- Agent-facing chatbots
Not convinced yet that automated customer service is the right fit for your company? Let’s take a closer look at some of the major benefits of embedding more sophisticated automation into your service toolkit:
1. Faster response times
Three-quarters of consumers say that speed is the single most important factor for a great customer service experience, according to Forrester. And the kind of speed consumers expect just isn’t feasible for most companies, at least not at scale. For example, world-class customer response time benchmarks for email, Facebook, and Twitter are all 15 minutes, and customers expect replies in under a minute on live chat, SMS, and phone calls.
With automation, companies can finally meet and even exceed these expectations. For instance, email automation technology can read incoming emails and auto-generate replies for agents to review and send in one-fifth of the processing time. In the diagram below, you can see the workflow process this technology follows to deliver fast, accurate responses. Learn more about how the email automation process works in this video.
Smart social listening software identifies urgent or highly emotional customer posts and use rules-based routing to get to the right agent immediately. And well-designed customer service chatbots can serve as the first line of defense on web and mobile channels, delivering specific and accurate resolutions within seconds.
2. Reduced customer friction
A study from Capterra found that 72% of customers blame their bad service experiences on having to explain their issue to multiple people. Automatically passing the correct information to the correct people minimizes this risk – not to mention, chatbots and agent-assistance automation can improve first contact resolution to the point where fewer transfers happen in the first place, doing away with the need for customers to repeat their issue.
72% of customers blame bad experiences on having to repeat their issue multiple times.
3. Expanded service hours and channels
Customers expect companies to be available and responsive on many more channels than they did even a decade ago. The only way to support new contact channels without adding significant headcount is through automation. For example, companies choosing to offer Facebook Messenger customer service often opt for a chatbot as the front line, escalating to a human agent only for those issues the chatbot is unable to solve.
4. Reduced cost per interaction
Self-service options like chatbots that can resolve issues before they ever reach a live agent represent a major cost-saving opportunity in part because they preclude a phone call, deflecting volume away from more expensive channels. Another example: using an email automation tool can reply to incoming customer emails for $1 per email, compared to the $5-6 average cost when a human agent is processing the same email (in some cases, we’ve heard this as high as $12 or $13!).
Call center automation means you can shift most expensive (human) resources to more meaningful work, allowing them the time to handle more nuanced customer issues as well as manage and improve the automation tools.
5. Higher agent engagement
It may seem counter-intuitive, but contact center automation tools actually benefit agents. By giving them the technology to speed through the mundane aspects of the work, they get to spend more time engaging with people (which is probably why they chose this field in the first place).
Higher engagement translates to lower turnover, which has a significant cost. IBM found that it takes an average of $4,000 to hire a new agent and at least another $4,800 to train them. With current contact center turnover rates averaging between 30-45%, that can add up to significant amounts.
It takes $4,000 to hire a new call center agent, and another $4,800+ to train them.
6. Fewer human errors
Call center automation technology helps reduce human errors. While people are great at empathizing with others, they aren’t able to match the accuracy and speed of automation when it comes to repetitive, tedious tasks like entering data into a case or crawling through thousands of knowledge base articles to find the correct one. If you’re leveraging contact center automation tools with AI and machine learning, these technologies actually get better and more accurate over time. This is due to their ability to learn from previous interactions, absorbing new information that changes their future behavior.
There are two main ways in which automating certain elements of customer service can future-proof your business: giving you the scale to handle massive volume spikes, like those seen, like those seen during the COVID-19 crisis in 2020, and allowing you to better engage with younger consumers.
During the early weeks of the COVID-19 pandemic, call centers experienced unprecedented contact volumes as panicked consumers sought information – and asked a whole set of brand-new questions, like when toilet paper and hand sanitizer would be stocked again, what precautions were being taken against the virus, whether store locations were open, and so on. Without automation in place, there is simply no way to handle drastic fluctuations in volume. For example, companies that weren’t using customer service email automation during this time saw their backlogs grow to the hundreds of thousands of unread, unanswered customer emails.
Another way in which automated customer service can future-proof your organization is by enabling Millennial and Gen Z consumers, who comprise more and more spending power, to engage in their preferred way. According to a study by Capterra, 70% of Millennial women report being comfortable solving an easy issue without talking to a human.
The majority of Millennial consumers prefer using automated customer service.
Even beyond that, 60% of Millennials also feel good about themselves and the company when they can resolve an issue on their own. So not only do younger generations feel comfortable with self-service, successful self-service interactions actually strengthen their bond with a brand.
Ready to take the next steps along your customer service automation journey? These best practices will keep you out of trouble:
1. Think of automation as a collaboration.
As we mentioned above, there are things automation and AI are great at, and there are things humans are great at. When building your plan for automated customer service, focus on where your human team and the automation can play to each others’ strengths.
This is similar to the idea of “cobots” in manufacturing and warehousing, where robots are designed to work alongside humans to optimize efficiency. A great example comes from the fulfillment centers for Box.com, an e-commerce retailer. They guarantee lightning-fast shipping, requiring their fulfillment operations to go from receiving an order to having it packed for shipping in 15 minutes or less. They found the robots and automated tools they employed were way faster than humans at finding and retrieving the right items from the stockroom, but struggled with packing items efficiently into boxes for shipping – there was too much empty space in each box and too many boxes were going out in a single shipment.
Turns out, no current technology can replicate the way humans pack items into boxes (maybe all that Tetris we played growing up?). So they redesigned their process to allow the robots to do what they did best, which was to locate the ordered items, and then pass them along to their human staff to pack the boxes. As a result, Box.com was able to hit their 15-minute shipping KPI.
For a customer service example, we can look to Astute’s Email Virtual Assistant (EVA). This call center automation technology automatically reads incoming customer emails, determines the correct CRM case coding by product and reason for contact, and generates a reply based on how human agents have previously responded to the same issue (including adding goodwill based on business rules). Agents review the reply, make changes as they see fit, and send. By incorporating agents in the QA process, you ensure any nuances that only a human eye would catch are covered before the message is sent. But because the reply is already generated based on previous agent replies, you don’t have to worry about the message coming off as robotic or unnatural.
2. Use automation only where it serves the customer journey.
For some touchpoints, human-to-human interactions will always be the best choice. For example, customers who are exceptionally angry or frustrated are usually best handled by a human agent. But in other cases, automation makes more sense. For instance, a conversational self-service chatbot can hunt down an estimated delivery date by combing through the order management system just as well, if not better, than a human can. Automation is best suited to tasks when it’s unnecessary or inefficient for a human to do that work.
What are some processes within your own customer service organization that could be done faster or better by machines? Here are a few examples:
- Automatically creating a case and populating reason code, product code, customer verbatim, suggested goodwill and responses based on previous interactions on live chat, self-service, social media, email etc.
- Scraping existing FAQ or knowledge base content to auto-generate chatbot replies
- Finding and routing high-priority, high-emotion social posts to an agent to reply immediately
The caution here is to avoid automating a process simply because it can be automated. Instead, determine what the benefit to the customer journey would be before proceeding.
3. Give customers an “escape hatch.”
Even the most well-configured, well-trained AI sometimes gets it wrong. No matter the interaction channel, always make live help available if your customers want it. According to research from Forrester, 60% of chatbots don’t give users the option to escalate to a human, causing considerable frustration when they feel the AI isn’t solving their issue.
Most chatbots don’t offer any options to escalate to a live person.
When a customer asks to speak to a live person, smooth the transition by keeping them in the current channel whenever possible and passing any relevant context to the agent, so they can pick up where the chatbot left off.
A common example is the self-service chatbot. When it doesn’t know the answer, many companies have designed the chatbot to direct the customer to their forum or knowledge base. While this certainly seems more convenient for the company, it’s a frustrating CX. If the customer is already chatting, albeit with a bot, they clearly want to continue chatting to find their answer. The best practice would be to immediately transition them to live chatting with a human agent.
4. Prioritize cross-channel consistency.
There are few things more frustrating than searching for an answer only to find the same company says different things in different places. Customers care about consistency!
Deliver this by using a unified platform for all customer touchpoints, regardless of channel, and drawing answers from a single base of information. The best practice for knowledge management is to use one knowledge base that’s able to tailor the length, format, and detail of its content to the channel, device, and user. For example, a customer engaging with a chatbot on your desktop website vs. Facebook Messenger might see content presented differently to suit different screen sizes. And an agent searching that same topic from within the CRM would see a deeper level of detail than what is shared publicly.
5. Start by automating the low-hanging fruit.
There are so many processes that are ripe for automation (no pun intended). But it’s easy to get overwhelmed by where to begin. As with any initiative, it’s helpful to map out which items will have the biggest impact for the lowest effort.
For example, enabling your new chatbot to answer the top five most common customer questions could make a massive impact for a relatively small time investment. But training the bot to answer every question your customers might ever think of? A huge effort for relatively little additional impact. Better to remember the 80/20 rule: 80% of incoming customer questions will be about the same 20% of topics, and focus your energy there.
6. Use tools already designed to automate elements of customer service.
While there are generic AI, machine learning, and RPA tools that can be used to automate customer service, it’s best practice to use a solution that already has automation built into their customer engagement toolset. There are so many challenges that are unique to customer service and consumer engagement, which makes working with a vendor with deep industry expertise a wiser choice. For example, Astute has embedded a variety of call center automation technologies into our customer engagement platform, saving you from having to recreate the wheel.
7. Gather feedback for continuous improvement.
Whether it’s through customer surveys or simple star ratings at the end of interactions, gathering customer feedback on automated processes – and acting on that feedback – will ensure you can continue to improve the customer journey. And you can even use more automation in that process! With advanced natural language processing and analytics, Astute’s chatbot platform can identify which questions are being asked that currently have no answer, grouping topics together to save you from having to analyze each individual customer question.
As with any promising new technology, there are some common pitfalls of customer support automation. Watch out for these five:
1. Automating for its own sake
Don’t get carried away automating processes for automation’s sake. Just because a process can be automated doesn’t mean it should be. For example, a customer with a complaint with the potential to become a lawsuit should be handled by a human, even if all the steps of the process could, in theory, be automated.
2. Not getting agents on-board
With AI and automation handling higher volumes of simpler service tasks, agents are increasingly tasked with only the complex cases. Forrester refers to this as the “automation paradox,” and it can make work more challenging for agents if they haven’t been provided with the right tools and training. Make sure you’re involving agents in both the planning and execution of new automation initiatives so that you can be better tuned-in to unforeseen challenges the process change might present for them.
3. Losing the personal touch
A study by Nextiva found that 33% of customers who stopped doing business with a company did do because experiences were not personalized. When we automate a process, there is a risk it can lose the personal touch. It’s important to look for ways to personalize experiences for customers. For example, could you…
- Personalize chatbot replies based on previous interactions with a logged-in user?
- Tailor product recommendations based on availability at stores near them?
- Automatically present the agent with the history of previous questions and complaints from the customer, allowing them to customize their response?
As you can see, automation technology can actually be used to increase personalization instead of decreasing it.
4. Going overboard on social media automation
Once you’ve experienced how chatbots can automate responses to common customer questions, it’s tempting to try a similar type of automation on social media channels. For example, wouldn’t it be great if the same chatbot could respond publicly to tweets and Facebook posts from customers?
As it turns out, this is not the right approach managing public post replies on social networks. Not only are these platforms cracking down on non-human activity, a substantial percentage of customers who post to your public social pages have already tried and failed to get a resolution through another channel, making them likely to be extra-frustrated and volatile. A canned-looking or non-human response just might push them over the edge!
Instead, employ automation to listen across social networks and surface the posts that need your attention, and automatically route them to a human agent for response. You can also embed an element of support automation within social by deploying a self-service chatbot on Facebook Messenger. Because it’s a messaging platform instead of a public forum, this channel is much better suited to a back-and-forth bot conversation.
5. Trying to “set it and forget it”
Automated customer service is not a Crock-Pot. Even though the word “automation” implies humans won’t need to be involved, the truth is that even the smartest AIs need human guidance. As you add more sophisticated automation to your customer service processes, think about who on your current team should be dedicated to implementing, monitoring, and enhancing its performance over time. And don’t forget that vendor partners like Astute offer access to Managed Services, a team of experts who can serve as an extension of your team to help you get the most value from the software.
Astute specializes in applying cutting-edge AI and automation technology to customer service’s biggest challenges, helping some of the world’s most influential consumer brands improve both contact center efficiency and customer engagement. Here are some of the tools we use to get it done:
- Award-winning customer service chatbot that delivers a quick ROI for service automation
- Email Virtual Assistant (EVA) that can cut customer email handle times by 5X or more
- Industry-leading agent desktop CRM with embedded robotic process automation and AI for efficiency
- Robust social media monitoring and routing to help you win at social customer care
- Conversational IVR that uses AI and natural language processing to reduce customer friction
Combining our product capabilities with decades of consumer engagement experience, Astute can help you reap the benefits of automating customer service while improving the customer journey.
Learn more about Astute’s customer care automation solutions. Schedule a live demo today.