Skip to main content

As with most buzzy new technologies, AI has many useful applications—and plenty of others that are gimmicks. But when it comes to customer service, the hype is justified: Gartner, an IT research company, estimates that chatbots will be the main customer service channel for 25% of companies by 2027.

Many companies are enthusiastically integrating AI-powered tools like chatbots and predictive analytics into their customer service processes. And rightfully so: this technology enables teams to resolve issues faster, and free up human agents to focus on customer relationships rather than answering the same question for the thousandth time.

Still, it’s worth getting specific about the benefits:

What KPIs are improved by adding AI into the customer service mix?

In this article, I’ll walk you through seven of the key performance indicators that are most impacted by adding AI into your customer service equation. By keeping an eye on these customer service metrics, you’ll be better positioned to use AI to reduce costs, increase customer satisfaction, and improve the customer experience.

What Are The Top Customer Service AI Tools?

Broadly speaking, AI customer service tools fall into three buckets: customer-facing chatbots and virtual agents, AI-powered analytics software, and back-office AI tools that empower support reps and automate processes.

Customer service teams often implement tools like:

  • Conversational Chatbots: These natural language processing bots can have text or voice conversations with customers to handle common customer queries without the need for human agents. Chatbots access knowledge bases to resolve routine issues like account information, product questions, and FAQs.
  • Virtual Agents: Think of virtual agents as a kind of “advanced chatbot” that makes use of deeper learning algorithms to understand context and hold longer conversations with customers; it’s also more likely to have a name and a personality. Like chatbots, virtual agents act as a first line of defense, resolving more complex issues before transferring to human agents as needed.
  • Predictive Analytics: Predictive analytics tools are powered by sophisticated machine learning models that analyze customer data to uncover patterns. Predictive analytics tools are especially useful for anticipating customer needs and helping agents personalize service; they can also be used to reduce churn.
  • Sentiment Analysis: Sentiment analysis tools that detect customers’ moods and emotions during conversations by analyzing text and voice data, allowing virtual agents to perceive dissatisfaction and adapt their approach accordingly—or escalate the conversation to human agents.

KPIs Improved By AI Customer Service Tools

The success of AI customer service tools is largely defined by how much of the customer service burden they can remove from human agents. Deflection rate, cost per resolution, and ticket volume all measure this factor directly or indirectly. 

Speed tends to be the other focus: metrics like first response time, average handle time, and first contact resolution can all track how quickly customers are being helped.

Boosting these support metrics also goes a long way toward improving key metrics like customer satisfaction score (CSAT), net promoter score (NPS), and customer effort score (CES).

Get the latest from the brightest minds in CX, UX, and design thinking.

Get the latest from the brightest minds in CX, UX, and design thinking.

  • No spam, just quality content. Your inbox is safe with us. For more details, review our Privacy Policy. We're protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
  • This field is for validation purposes and should be left unchanged.

1. Deflection Rate

An increasing deflection rate is one of the clearest signs that your AI customer service investment is paying off. If you’ve invested in chatbots, virtual assistants, or other AI-powered self-service tools, your deflection rate—the percentage of chats that your artificial intelligence systems handle without the support of human reps—should be substantial.

Fin, an AI chatbot designed by Intercom, boasts a deflection rate of up to 60%. Depending on the complexity of your support inquiries, some chatbots can handle as many as 80% of inquiries without human intervention.

Tools like Fin manage this feat through a combination of natural language processing, pre-defined conversation flows, and a deep understanding of your product documentation. The end goal is for the most common support inquiries—like account information, order status, returns, and simple troubleshooting—to run entirely through AI chatbots, freeing up human support agents to focus their attention on complex inquiries and relationship-building.

2. Cost Per Resolution

If you look at worldwide customer support activity in aggregate, the total numbers are staggering: each year, customers submit 265 billion support requests. Chatbots and AI assistants help offload some of this expense, reducing the burden on human customer support teams and saving businesses more than $11 billion per year.

AI’s impact on cost per resolution comes down to one factor: it’s cheaper than humans. That means that as you shift more customer inquiries to self-service AI, your average cost per resolution should decrease.

So what precisely, in dollar terms, is AI’s impact on cost per resolution? Making generalizations about the cost of a customer support interaction is notoriously hard to do across companies. There are too many variables: the industry, complexity, and cost of labor, just to name a few. But to give you a sense of the possible savings, LiveAgent, a customer support tool, says the industry average for cost per contact is $7.16. Meanwhile, Fin—the aforementioned AI chatbot designed by Intercom—is 86% less at a cost of $.99 per resolution.

3. First Response Time (FRT)

Savvy CX leaders may be wondering where their focus should lie in the age of AI. One classic customer experience priority is more relevant than ever: saving customers’ time. According to Forrester, 66% of customers say the most important thing a company can do is value their time. As chatbots have become widely used, customer expectations have risen: 40% of customers expect a response from a chatbot within five seconds.

AI is nothing if not fast: chatbots can respond to customers within a few seconds, 24 hours per day. For customers who need a human rep, virtual agents can gather information and get a live customer service agent on the line within about 30 seconds. From there, AI can help further by providing agents with context-aware suggestions.

The latest help desk software—enhanced with AI—makes it easy to watch all of this chatbot and human activity from a bird’s-eye view, automatically applying analytics so you can see trends and opportunities for process improvement.

4. Ticket Volume

One of the easiest ways to gauge your overall customer support activity is to look at ticket volume. The surface-level implications are obvious: high numbers mean a busy team. 

But ticket volume is also connected to broader indicators that tie into customer happiness. One survey found that companies with lower ticket volume tend to have higher CSAT scores. As support ticket volume goes up, customer support teams get stretched thin and as a result, CSAT scores tend to go down.

Fortunately, AI tools are well-equipped to reduce ticket volume. Chatbots are the primary method: they can deflect many customer inquiries before they ever turn into tickets. Having a robust knowledge base—which is then used to train your virtual agents—is another key part of the equation.

5. Average Handle Time (AHT)

The ideal average handle time for one business isn’t necessarily relevant to another; low complexity businesses like retail tend to have a relatively low AHT, for example, while telecom companies skew higher. But in general, a “good” AHT is around six minutes in a traditional call center setting.

Live chat tends to have longer AHTs because both agents and customers can multitask, with agents often handling three or more chats at once. Still, if a customer reaches out with a problem, the faster it’s solved, the happier they’ll be.

For simple queries, chatbots can speed up AHT by providing the answers to many common questions nearly instantly. For more complex issues, virtual agents can quickly collect the necessary information from customers before passing them off to human agents to take the final steps.

6. First Contact Resolution (FCR)

There’s nothing more frustrating than contacting customer support, only to wind up reaching out again because the issue wasn’t solved the first time. By the time customers reach out for the second time, the lack of a quick resolution may have already damaged their perception of your company’s service.

That’s why it’s so critical to resolve issues the first time customers contact you. Ideally, four out of five customer issues should be resolved upon first contact: 70-79% is considered a ‘good’ FCR rate, according to contact center industry benchmarking conducted by SQM Group.

AI-powered virtual agents can improve your first contact resolution rate by helping customers resolve their issues quickly through self-service. AI can also help human agents resolve more issues upon first contact by summarizing past customer interactions and suggesting solutions.

7. Retention Rate

There’s no more important metric for SaaS companies than retention rate. Fortunately, AI can help: AI-powered predictive analytics and sentiment analysis tools can crunch vast quantities of customer data to uncover signals that indicate when a customer is at risk of churn.

Data like demographics, purchase history, usage history, sentiment, and satisfaction surveys can be hard to make sense of on their own, but analytics tools can detect subtle patterns and create unique churn prediction models to proactively win customers back before they leave.

The Power Of AI-Enabled Customer Service

AI has the potential to produce a rare “win-win-win” in the world of customer service: lower costs for companies, faster service for customers, and less repetitive work for support reps.

A virtuous cycle links these three factors together. By delegating high-volume customer service inquiries to AI, businesses reduce costs and human agents are freed up for work that plays to their strengths. Customers get a combination of faster service for routine interactions and more empathetic service for complex interactions, improving retention and satisfaction.

By tracking the right customer service KPIs, you can ensure AI is having the desired effect on your customer support experience—creating happier, more loyal customers.

For more on KPIs, check out our in-depth article on the top customer success metrics. Want to stay on top of the latest in the CX industry? Make sure to subscribe to our newsletter for CX leadership tips, marketing strategies, insights, and industry trends.

By Ryan Kane

Ryan Kane has been researching, writing about and improving customer experiences for much of his career and in a wide variety of B2B and B2C contexts, from tech startups and agencies to a manufacturer for Fortune 500 clients.