When AI Meets CX: The Future of Customer Service
Table of Contents
Artificial Intelligence (AI) is all the rage. The field has seen pretty intense growth in the past years. Recently, businesses have also started using AI to increase efficiency and deliver services at scale—and yes, this includes customer service.
With all the conversations about AI, these questions are popping up more often too:
‘What’s next?’
‘Where is this going?’
‘Will AI really replace customer support agents?’
While we don’t have a crystal ball and superpowers, what we did was consolidate insights from our network of partners. In this guide, we dig deep into the development of AI’s role in customer service, and where we think this is headed.
How has AI transformed the customer service industry?
Basic chatbots started with handling Level 1 requests
Chatbots have been a popular starting point for businesses wanting to integrate AI into their customer support operations. The early versions were designed to handle minor requests like providing basic information or directing customers to appropriate resources.
GQ Fu, CEO and Co-Founder of LTVplus, explains that the use of chatbots gradually accelerated. Though capabilities were limited, the chatbots were still able to provide some form of valuable support to customers 24/7.
“We got basic chatbots in the first few years,” GQ says. “But when you talk about having more intelligent solutions like the AI itself chatting with customers, we didn’t have that at the time. To have that, you had to have a lot of data sets for training the AI—so you had to spend a lot.”
Jen Staben, Director of Customer Experience at Tydo, loves the concept of AI handling minor requests to free up agents so they can focus on doing what they do best – building human connections.
AI assistants entered the scene but they still required a lot of training
After chatbots came AI assistants that were all the rage. After all, they offered faster and easier ways of doing things.
However, AI assistants needed a lot of training. Essentially, the performance of AI assistants was very dependent on how they were set up and “trained” by humans.
“It’s not like you switch it on and AI’s working, yay! It’s a journey—stack by stack,” explains Stephen Jones, Head of Partnerships at DigitalGenius.
Stephen also outlines the AI evolution of DigitalGenius.
- DigitalGenius started out as a custom-built solution for SMS replies for a large car manufacturer.
- The AI was then used to suggest potential replies.
- Intent Detection followed as well as template suggestions for agents.
- Over time, the templated responses became so accurate that the AI was allowed to reply to respond directly to customers.
- Automation of processes followed, such as processing refunds, arranging replacements, tracking delivery status, and even passing information to other team members or departments.
AI tools evolved to provide more personalized experiences
The growth of AI continued. Pretty soon, there were a lot of new developments that had a significant impact on customer service:
- Natural Language Processing
- Machine Learning algorithms
- Sentiment Analysis
Sentiment analysis, in particular, has been an ultimate game-changer. By helping identify customer emotions, the customer support team becomes equipped with information that will help them improve the overall customer experience.
Here’s an example:
Neil Forrest, Sr. Strategic Partner EMEA at Gorgias, shares how they were able to gather data from over a hundred million tickets a year. They identified that a third of tickets are from the pre-sales stage. Because of this information, customer service teams prioritized improving the quality of pre-sales support.
This doesn’t mean ignoring post-sales support tickets though. Enyrck Serin, GTM Lead Partnerships at PolyAI, brings up a good point.
Discover the difference that conversion-websites make.
*LTVplus customers can use a special code for 30% off. Ask us about it!
Businesses got really excited at first, but only a few pushed through
The AI hype was definitely real. Everyone wanted to jump in right away because they didn’t want to get left behind. However, brands soon realized that AI wasn’t simply plug-and-play.
“It took a lot more work than people liked,” GQ explains. “That’s why brands didn’t adopt. Some did, but only to a certain degree. Plus the costs were high at that time.”
Maurits Pieper, Head of Partnerships at Dixa offers additional insights. “The tech was never fully up to par then. People got really excited at what it can do, then they got slightly disappointed when they set it up.”
Additionally, when businesses started to test out AI, they started to feel hesitant about handing over critical pieces of customer service to AI. “What happens when you actually hand over a significant part of your business to AI? What does the transition period look like?” Maurits expounds.
So while businesses were excited, they also realized that it wasn’t that simple.
3 ways that online businesses are using AI to enhance the customer experience
Optimizing help articles and knowledge base
Jen Staben gives a reminder that customer service also includes help articles and knowledge bases.
- FAQs
- Help articles
- Knowledge base
“A lot of people don’t really want to talk to chatbots or agents,” she explains. Contacting support is usually a last resort if customers cannot find the answers elsewhere. However, comprehensive help articles, FAQs, and knowledge bases are difficult to create—not to mention time-consuming.
With tools like ChatGPT (which is all the rage today), businesses can save a lot of time crafting these resources.
Crafting efficient and effective responses
Aside from crafting help articles, guides, FAQs, and knowledge bases, AI tools also improve the efficiency of the responses to customers. At the same time, the human touch is still there. The AI tools suggest responses and churn them out quickly, but it’s still the human agents who deliver the message.
“Another trend worth mentioning is the use of predictive analytics to anticipate customer needs and offer proactive support,” Tshili Khupe adds. “This reduces the likelihood of customer churn.”
Providing personalized product recommendations and cross-selling
Essentially, recent developments in AI have paved the way for not only faster customer service—but also more personalized.
Neil explains that AI has been absolutely fundamental to the success of Gorgias and their clients. Here are some concrete examples:
- Identifying sentiments in organic social and paid social
- Identifying negative comments so they can be prioritized
- Faster responses and engagement in posts
- Faster identification of post-purchase questions
- Monitoring order tracking, shipping policies, and faulty products
A more personalized approach increases customer retention—which is going to be a very big focus this year according to Sara Pereda, Senior Partner Manager at Yotpo.
Sara explains that businesses have to keep their customers engaged so they will keep coming back. While one-to-one interactions are ideal, it’s hard to implement this at scale.
Meanwhile, AI helps offer the right products to the right people at the right time. AI can also personalize product recommendations and cross-selling. Equipped with this information, customer support teams can facilitate higher-quality conversations that will convert.
Meet our modern superheroes: How can AI assistants swoop in and help customer service agents deliver better experiences?
Agents can deliver faster responses thanks to automation
Guillaume Luccisano, Founder of Yuma, outlines a number of current possibilities thanks to automation:
- Automated knowledge base resource highlighting
- Brand voice imitation layer on each message sent by an agent
- Summarization of threads or conversations for automated escalation
- Automated drafting of suggested answers
- Automated workflow that is also entirely translated
- Drafting of a complete message and response
- Intention detection
- Automated extraction of data for trends and reports
While human customer support agents are more than capable of these tasks, the speed at which AI tools can accomplish them is significantly much faster.
Start building better conversations.
AI tools can handle the repetitive work so agents can focus on complex tickets
In customer service, it is expected that a lot of inquiries and tickets will be similar. That’s why businesses have FAQs, knowledge bases, and corresponding scripts for human agents to guide customers through resolutions.
As AI tools become more advanced, they can start handling simple and repetitive support tickets. Stephen elaborates on this. “AI-powered tools can deflect tickets so humans don’t have to reply to every single ticket. This will then free up their capacity to focus on more complex stuff that AI cannot solve—yet.”
One important thing to note: handing off tasks to AI does not mean that human customer support agents will be redundant.
AI tools can provide real-time insights
The reason why human customer support agents take longer to respond is that they are still trying to analyze the customers’ sentiments.
Of course, the more unique and complicated the support ticket is, the longer it will take to understand it and respond. The process usually involves keyword tagging and then sentiment analysis based on specific criteria.
This is a massive opportunity to involve AI. Maurits explains that AI tools can take care of keyword tagging and sentiment analysis on a surface level. That way, agents can already get a preliminary understanding and be able to respond faster.
At the same time, AI tools can help with the reporting process to identify recurring issues and trends in real time. There might be some key issues that sales, marketing, and even product development teams might need to know about.
4 challenges that companies face when integrating AI assistants in customer service
Lack of resources needed to set up, implement, and manage AI assistants
As mentioned above, AI is not plug-and-play.
It still requires a significant amount of resources. Aside from being costly, key team members must spend a lot of time briefing and monitoring the AI. Additionally, even when you reach the point wherein AI is already handling a percentage of your customer service department’s tasks, you will still need a dedicated resource to ensure that you are maximizing its value.
Maurits couldn’t explain it any better. “It would be great to have AI handle 40-50% of repetitive inquiries. If you have a lot of these repetitive tickets, you have to maximize the value you get from these interactions.”
Realizing that expectations were too high
With the hype around AI right now, it’s easy to be disappointed when it doesn’t do something the way you expect it to.
Business owners need to be constantly reminded that AI is only as good as the technology and the brief it receives.
Stephen gives a great example. “The ChatGPT hype is real and so are the issues surrounding it. In a week there are already so many tools that are integrated with it. However, you can’t let AI answer your customer just yet. Let’s take Nike as an example. If customers ask what the best running shoes are, the AI can answer ‘Adidas.’ You lose all control if you relinquish everything to ChatGPT and you need restrictions in place.”
“Expectations from companies are too high after seeing ChatGPT,” Guillaume agrees. “But technology is not fully ready yet for complete adaptability to a specific business.”
Essentially, some businesses tend to expect that once they install AI, it will be good to go. According to Guillaume, if you choose to automate these repetitive tasks, you will need the AI to be 99% right because you do not want errors to affect the quality of the customer service you’re delivering.
Handling the initial reactions of team members
There will always be the possible pushback from the team if you decide to adopt AI. People will instantly think that they are going to be replaced by robots and that they will lose their jobs.
In the customer service industry, for example, agents might be reluctant as they fear for their role.
To avoid this, it’s important for business owners to approach the matter strategically by helping agents understand what AI will do and how it can help them.
Restructuring the team and the transition period involved
Related to the issue above, the entire transition period can be rough—from handling initial conversations with the team to increasing AI’s involvement.
GQ has some advice to business owners:
- Understand the direction of your business moving forward so you can communicate this effectively
Come up with the short-term and long-term objectives of integrating AI
Plot out the target phases and timelines so the team gets the bigger picture
- Lay out the structure of your Customer Support department with AI included
Are the agents briefed on the AI tool and it’s features?
What is the AI tool expected to do for your business?
What will the agents do with their free time once the AI has been onboarded?
What will be the changes in the agents’ KPIs and expected output?
Invest in delivering an awesome customer experience
Blake Imperl, Head of Merchant Growth & Enablement at Wonderment could not emphasize it more. “Invest in customer experience. Double down on your transactional channels and study the experience once a customer makes a purchase.”
3 Strategies to maximize AI assistants and customer agents
Prioritize human-to-human interactions pre-sale
As Neil mentioned earlier, a lot of support tickets come through in the pre-sales phase. What does this mean? This means that there are a lot of revenue opportunities there, and business owners should prioritize human interactions at this stage.
Empower your agents by educating them on how AI can help
As Guillaume said, “AI is here to be an assistant for the agent—to give them superpowers.” To truly maximize the powerful combination of human customer support agents and AI, you will need the agents to back this up.
The way to get your agents to truly embrace what AI can do for them is to empower them. It’s important for them to know how AI can help them, what the limitations of AI currently are, and how they should go about the transition.
Jen emphasizes that customers do not usually want to have to contact customer support. They reach out when a help article, a FAQ section, or a knowledge base doesn’t make sense and doesn’t provide a solution. “So how can this push us forward to come up with a solution so that there’s no problem in the first place?” she asks.
The answer lies in the successful empowerment of agents, such that they are able to take the data from the AI and translate these into concrete actionables that will help improve the overall customer experience—so that there are no problems, to begin with.
Integrate AI assistants in phases
Integrate AI in your process in phases. Start by automating a lower percentage of interactions and slowly ramping up as needed.
Maurits’ advice is to start small at 10-15% so you can test it out while avoiding internal friction. After a few weeks, you can gradually increase to 20-25% until you eventually get to a point wherein you are maximizing AI.
Tshili explains that it’s important to always stick to the primary goal of implementing AI in CX. “It’s to improve the efficiency of communication, speeding things up, making customer interactions quicker and straight to the point. I don’t think it will fully take over customer service.”
Invest in delivering an awesome customer experience
Blake Imperl, Head of Merchant Growth & Enablement at Wonderment could not emphasize it more. “Invest in customer experience. Double down on your transactional channels and study the experience once a customer makes a purchase.”
AI + human customer support agents: three years from now
It’s safe to say, and we have to establish it now—that AI is not going anywhere. By the year 2030, it’s expected that AI will contribute $15.7 trillion to the entire global economy—that’s 26% of the global GDP!
But let’s look at something a little closer…. How about three years from now?
Prediction #1: Widespread adoption of AI-powered tools
According to Tshili, advancements in natural language processing and sentiment analysis will continue and the adoption of AI-powered tools will be more widespread. However, this doesn’t mean that human agents will be wiped out—it’s the opposite actually.
“It’s important to note that while AI will play an increasingly important role in customer service, human agents will still be necessary for complex and high-touch interactions,” Tshili explains. “If I were to sneeze on a call, would the AI ‘bless me’?’”
So while AI’s role in customer service will indeed flourish, companies must ensure that the responsible development and deployment of the technology is a top priority.
Prediction #2: There will be more voice assistants with automation at scale
This is Enryck’s prediction—more conversational assistants with automation at scale. He explains that AI will be powerful enough to understand customers, but that humans will need to step in depending on specific use cases. However, AI voice assistants with automation will make the process more efficient, allowing human agents to respond much faster.
“Automatic workflows will give agents more time. There are very few situations where you want AI to be creative—so you will still want humans to handle those scenarios,” he adds. “It’s all about the balance between what the AI and human agents will do.”
Enhance your CX with voice AI
Prediction #3: Different AI integrations will work together
Neil would eventually love to see a brand experiment and automate 100% of tickets in a day just to see the impact on the business and customer. Of course, this is still too risky to do today, but it might be possible in three years’ time.
He also predicts that “investing in AI” will not mean investing in just one tool. Different AI tools and integrations will work together, with image detection becoming a big part of eCommerce customer support.
IMAGINE THESE SCENARIOS
Scenario #1
A customer snaps a photo of a faulty product. The AI detects and identifies the damage and proceeds to process the refund
Scenario #2
AI assistants will process returns and transactions while human agents are doing other things—allowing multitasking at scale.
Prediction #4: Every single hole that can be filled will be filled by AI
Jen shares that of course, AI will be here to stay, but in three years we may be in a position wherein a customer calls or chats with a company any time of the day and there’s a response from AI.
“Every single hole that can be filled will be filled with AI,” she says, expressing that there will probably be an AI tool for every little thing. Need to do this? There’s a tool for that. What about this? There’s also a tool for that!
So what’s the issue with this? Check out the next prediction.
Prediction #5: Unique strengths and abilities of human customer support agents will become more important
“AI and automation are put in place to help the workers be better, but what happens if it helps the worker become obsolete?” Jen asks.
Jen explains that if there is an AI that will tell customers what to do each and every time, this setup might be more cost-effective than forwarding a support ticket to engineers or tech people.
In the next three years, valuing and maximizing the unique strengths of human customer support agents will become increasingly important.
While AI-powered tools will continue to play a vital role in enhancing the customer experience, businesses will also prioritize empowering their human agents to perform at their very best.
This may involve the following:
- investing in training programs
- providing agents with AI-powered tools that augment their skills
- creating a company culture that emphasizes the value of human interaction in customer support.
As AI becomes more widespread, businesses that prioritize the human touch in customer support will stand out in a crowded market.
Key takeaway: AI will never replace a human, so it’s time to put a premium on what humans can do that AI cannot, instead of purely focusing on what AI can do faster. What is it that human agents can do more to improve customer experience, with AI handling the other tasks?
Contributors
- GQ Fu, CEO, LTVplus
- Tshili Khupe, Head of Business Development & Partnerships, LTVplus
- Stephen Jones, Head of Partnerships, DigitalGenius
- Maurits Pieper, Head of Partnerships, Dixa
- Neil Forrest, Sr. Strategic Partner Manager EMEA, Gorgias
- Enryck Serin, GTM Lead Partnership, PolyAI
- Jen Staben, Director of Customer Experience, Tydo
- Sara Pereda, Senior Partner Manager, Yotpo
- Guillaume Luccisano, Founder, Yuma
Increase sales and customer satisfaction today.
Build a dedicated customer service team that supports the growth of your brand.
At LTVplus we take care of recruitment, training, and optimization for you.