There have been vast improvements in artificial intelligence (AI) in recent years.

So much so that AI implementation is not only top of mind for almost all organizations, but has become the new frontier for business competition on everything from AI-improved employee performance management to marketing and growth analytics.

In customer experience and customer care, the case for AI is crystal clear.

Chatbots are no longer "mostly bad" at their intended job, and for far too long costly customer service agents have been doing tedious, repetitive tasks that AI specializes in. B2C organizations have hastly implemented AI in customer care in particular, in an effort to reduce handling costs, triage tickets efficiently, and extract better insights to improve CX.

In this article, we'll focus on a use case that's growing quickly as a priority for CX teams: AI customer analytics.

AI is brilliantly successful at unlocking a world of CX insights that were previously stuck in arduous data black boxes.

Whether you're looking to improve your tagging, understand surveys at scale, or drive CX outcomes across your organization, AI customer analytics may be the innovation your team needs to invest in in 2024.

In this article:

• What Is A Customer Service Analytics AI?
• How Does a Customer Service Analytics AI Work?
• Three Types of Natural Language Processing AI for Customer Service
• Why does AI work particularly well in customer service?
• How To Use Customer Service Analytics AI In Your Business

What is AI Customer Analytics?

AI customer analytics is the process of pulling valuable insights from our customer touchpoint data using analytics techniques like sentiment and text analysis.

For most companies, the majority of customer data comes from two areas:

1. Customer feedback: surveys, customer reviews, social media.
2. Customer service logs: calls, live chats, emails.

The latter, customer service logs, are easily your most powerful ally—they have less bias, they're frequent and high volume, and they don't rely on customer memory to give an accurate account of the customer experience.

However, most companies fail to capitalize on customer service conversations because they'e difficult to analyze and extract insights from due to their complexity and scale.

AI customer analytics thrives in this environment. Machine learning-based sentiment and text analysis tools, like Sentisum, make it easy to:

• Uncover the real drivers of customer friction
• Attach sentiment to topics of conversation
• Identify patterns and predict trends over time

AI customer analytics works through what we call "tags". Tags are groupings or categories, which can be high level like "refund", "delivery issues", "negative sentiment" or specific like "mouldy bread caused refund".

AI reads and understands feedback and conversational data and applies these tags, which helps turn qualitative text data in something quantifiable.

A bonus from having detailed tags applied across customer conversations, is the ability to setup rules to route, triage and prioritize inbound customer service conversations.

For example, the holiday rental giant, James Villas, used AI customer analytics to identify desperate customers with high-priority issues and routed those customer service conversations to an "urgent" queue in help desk provider, Zendesk.

Unlike a manual analysis, AI analytics is infinitely scalable, granular, and uniform in its approach to analyzing customer text and voice channels.

I asked Kirsty Pinner, Chief Product Officer at SentiSum, why AI is such a powerful tool in customer service.

Here’s Kirsty’s answer:

“AI can cut through the subjectivity of human opinion, and no matter how something is said, it can report on the customer issue in a simple way. The latest developments in AI analytics can handle complexity extremely well.”

“No other method gives a representation of customer conversations this accurately. Manual tagging is too subjective and trying keyword analysis is too blunt a tool.”

How Does a Customer Service Analytics AI Work?

So, how does it do it? And can it really be as accurate as a human? Haven’t we all heard that AI is still a bit rubbish really?

The main goal of AI is to simulate human intelligence, enhanced with the capabilities of a machine—infinitely scalable, never tiring, and doing it the same every time.

Put to use in a business environment, AI is particularly good at automating repetitive processes related to understanding, problem-solving and learning.

Two advanced subfields of AI are machine learning (ML) and natural language processing (NLP). These are built around algorithms, which take large datasets (e.g. the last 12 months of your support conversation logs) as their ‘training data’ and some initial human direction and together they understand the logic between words, sentences and phrases in the human language, learn from it, and can apply the logic consistently into the future like a human would.

In our article, ‘Why is machine learning NLP so good at Zendesk ticket categorization?’ we discuss the different types of Natural Language Processing AI in depth.

There are three key types of NLP, each of which produce a varied results. Let's look at all three in a customer service analytics environment.

Three Types of Natural Language Processing AI for Customer Service

If you have 10,000 monthly support calls, emails, chats, Messengers, WhatsApps and so on…the last thing you want to do is spend a week going through them manually. In fact, that’d be very boring and the results would probably be inaccurate.

A computer, on the other hand, doesn’t get bored and has no limits to its reading speed. It can process and categorize millions of conversations in near real-time with NLP.

Of the three types of NLP below, we recommend avoiding 1 and 2. Customer service leaders come to us all the time after trying them and feeling like their time is wasted.

Let’s look at why:

1. Keyword Extraction

You may have come across automated tagging in a helpdesk tool like Zendesk or Intercom. They rely on keyword extraction to tell the automation which tags to apply.

How does it work? Simple, if a keyword is present the follow a certain rule: 

“If the word ‘refund’ is found anywhere in the conversation → Apply the tag ‘refund_request’ to the conversation.”

There are obvious flaws in this approach. The main one being that there are hundreds of ways to ask for a refund without using the word ‘refund’, meaning that unless you can think of them all and tell the keyword extraction tool, your analytics will not have accurate results.

This approach fails to provide granular results, too, because it’s simplistic. Unlike other approaches, keyword extraction isn’t an intelligent form of AI, so it can’t tell you additional information like the why behind the request. You may know you had 1,000 mentions of ‘payment issue’ but not that 500 payment issues were caused by Klarna, 300 by Payal, and 200 due to a website bug.

2. Rule-based NLP

Rule-based NLP algorithms leverage libraries of ‘rules’ to help them understand human language.

Instead of you directly telling the automation what to do, rule-based NLP understands things like ‘liberty’ and ‘freedom’ are synonyms.

This has much less complexity and much more accuracy compared to keyword extraction. It has two downsides:

• If the rule doesn’t exist then it won’t understand the meaning.
• It’s still looking for words or sentences, rather than understanding the language like a human would. If a customer said “I’m not happy because my payment keeps getting rejected, I think it’s Paypal’s fault but it could be Klarna”, and then the conversation between the agent and customer continued to determine it was neither Paypal or Klarna, the rule-based NLP would still tag that conversation with “payment_issue”, “klarna”, and “paypal”.

In short, it still isn’t intelligent enough to provide meaningful insights.

3. Machine-learning based NLP

Machine-learning based NLP understands speech and text in a similar way to humans. After digesting a dataset being trained in it, it uses statistical inference to carry that knowledge into new environments it’s never seen before.

For example, it can identify and infer the meaning of misspellings, omitted words, and new words like slang by itself.

Thanks to the ML, it learns the patterns between phrases and sentences and constantly optimizes itself to improve accuracy over time.

We highly recommend you ask your provider the type of NLP they’re using and whether they will build you a customized model or not—if they don’t plan to, it’s unlikely to provide the accuracy or granularity you need to achieve your goals.

‍Deep Dive: Machine learning vs keyword tagging: Why ML is best

‍

How We Apply AI in Customer Service Analytics

There are two ways we can apply AI to customer service analytics: topic analysis and sentiment analysis.

The topic analysis method assigns topic tags or categories to text based on the underlying meaning, reason for contact or theme.

The example in the image above is a topic analysis applied to customer service tickets or survey free text. Topic analysis can go beyond the topic and subtopic of conversation to label things like intent and urgency.

On the other hand, a sentiment analysis simply determines whether human speech or text is positive, neutral or negative. 

Working together, these techniques form what we call a topic-based sentiment analysis. For customer service teams, this is your holy grail of insights.

If you run a topic-based sentiment analysis within your support department (using a tool like SentiSum), you’ll know in real-time the key topics driving survey results or support contact AND the sentiment attached to those topics.

For example, if you identify 50 topics that drive 99.5% of customer contact, you might not be sure which issue to fix first. Your first instinct might be to fix the highest frequency first, that would reduce the most amount of costs, right? 

But, with a sentiment analysis alongside your topic analysis you might discover that your highest frequency issues only cause a low level of discomfort for customers. You might also find that another topic only impacts a handful of customers, but it induces severe anger—driving negative reviews and bad word of mouth.

With all the insight at hand, you can make a decision that best suits your goals and resources at the moment.

See this in action in the SentiSum platform video on our homepage.

Why Does AI Work Particularly Well in Customer Service?

AI is only as good as its training data. Luckily, customer service teams usually have tons and tons of conversations the AI can train on.

Most support teams have thousands (if not hundreds of thousands) of conversations, calls, texts and historical surveys—all of which can be used to build and train a topic-based AI algorithm in a matter of days. 

With our customers we go through this process:

• Identify your priorities—what issues do you want to track? What goals do you want to achieve?
• Manually analyze and tag a few hundred or thousands of tickets—identifying patterns, common tags, and building a custom AI that understands the basics.
• Running the AI through all the historical ticket logs—and then optimizing away any inaccuracies.
• Use our proprietary tools to continuously monitor the AI and make sure the machine learning is working as it should. For example, if a brand new topic arises, did it catch it?

Thanks to this method of building custom algorithms from scratch, the AIs we build are incredibly accurate at uncovering topics and sentiments from customer service interactions.

That’s the power of AI—if it’s high frequency, complex and unstructured free-text, then the latest AI innovations can draw insights out of it faster and more accurately than anything else.

How to Use Customer Service Analytics AI in Your Business

The results of a customer service analytics AI create lots of opportunities for improvements in operational efficiency, customer experience and agent experience.

1. Make customer insights widely accessible

One of the most important ways to use AI in customer service analytics is to make the insights widely available.

Because AI is so accurate, granular and consistent in its analytical approach, we’ve seen that other departments across the company are quickly interested in the insights.

This LinkedIn post by Charlotte Lynch sums it up nicely:

"when every employee is able to access VoC data, they are empowered to set objectives and KPIs based on real-life numbers."

The more available you can make insights to the rest of the business, the better and more customer-centric everyone’s decision-making becomes.

Due to the real-time nature and high frequency of customer service interactions, paired with an AI analysis tool they can become the beating heart of the company’s decision-making.

That’s what happened at Gousto (full interview here) when they implemented AI analytics into their support function.

“We use the latest technology to give open access to our voice of the customer data across the business for teams to self-serve insights for anything from discovery work for Tech initiatives through to root cause analysis for any operational complaints to guide improvements.” Joe Quinlivan, Head of Customer Service at Gousto

As one of the most innovative, data-led British corporations to spring up in the last 5 years, it’s worth watching how Gousto does things.

2. Improve customer experience

AI helps you do more regular, in-depth customer service root cause analyses. As you’ll already know these are incredibly important for improving customer experience.

Here’s x ways you can use an AI-driven customer service analysis to improve CX:

• Create a knowledge centre to tackle FAQs
• Understand negative CSAT or NPS drivers and tackle them
• Use AI-tags to provide better customer service (e.g. prioritize urgent support tickets)
• Use AI tags to help you scale efficiently

We get really excited about this topic because the potential for improvement and time-saving is HUGE. Check out our article on the ‘Importance of Customer Service Analytics’ for more examples or our article on ticket tagging use cases here.

3. Do call centre predictive analytics

The future of customer service analytics is predictive. Imagine if based on certain topics or sentiments you could look to the future, know what’s coming and provide preventative solutions.

Call centre predictive analytics is an emerging field and the use cases are still forming. Here are a couple of examples we deploy with our customers right now:

• Link CSAT scores to conversation topics to understand what words or phrases used by agents is the best approach to solving specific queries. If the data shows that certain solutions or approaches lead to better sentiment and scores, then agent performance can be improved across the board by training them on that approach in the future.
• Predict customer churn risks. We developed a tag called “churn_risk” that understood when a customer contacted support about a certain set of topics, and mapped that with other predictive characteristics like sentiment, the AI algorithm then applied this tag to incoming conversations in real-time. As a high priority for the business and customer retention, these tickets could then trigger automated triage to the urgent queue, ensuring those requests were handled quickly and with care.

Leverage SentiSum as Your AI Customer Analytics

SentiSum customers use our AI platform to automatically surface important topics and customer sentiments from customer service conversations. With our accurate, granular tags you can unlock time-saving automations and experience-boosting insights in real-time. 

‍Book a product tour with us here to see how we can help you.