When people talk about great customer service, they usually jump to one metric: CSAT . And while CSAT is important, it’s just one piece of the puzzle . If you want to really understand—and improve—the customer experience, you have to look at more than just satisfaction in isolation.

That’s where CES (customer effort score), NPS (net promoter score), and AHT (average handle time) come in. These three metrics often get less airtime, but they offer deeper insight into loyalty, frustration, and operational efficiency—the ingredients of long-term success.

Think of them like a backstage pass to your support performance; CSAT shows you how customers felt after the interaction, while these metrics show you why.

CES, NPS, and AHT are key performance metrics for customer service, and they deserve a little more time in the limelight. Let’s break each one down, explain why it matters, and show you how to make them actionable in your AI customer service strategy .

1. net promoter score (NPS): the loyalty litmus test

what is NPS?

Net promoter score measures how likely your customers are to recommend your company, product, or service to others.

CSAT shows you how customers feel about specific interactions — and when it trends down, it’s often an early warning sign of bigger problems ahead. NPS, on the other hand, gives you a broader view of long-term loyalty and advocacy. High NPS means customers are more likely to renew, refer, and advocate—all critical to sustainable growth. And because it’s asked outside of a specific ticket context, it reflects the whole experience, not just one conversation.

NPS based on a simple question: “On a scale from 0 to 10, how likely are you to recommend our company/product/service to a friend or colleague?” Based on the score, customers fall into three buckets:

• Promoters (9–10): Loyal and enthusiastic. They’ll refer you to others and keep coming back.
• Passives (7–8): Satisfied but unenthusiastic. They’re at risk if a competitor comes knocking.
• Detractors (0–6): Unhappy and vocal. They’re likely to churn, complain, or damage your reputation.

why does it matter?

Of all the customer service metrics in your arsenal, NPS tells you the most about long-term customer loyalty. It’s not only a measure of how happy someone was after a single conversation, but it tells you how likely they are to stick with you and advocate for your brand.

NPS is your best signal for long-term loyalty—and a strong reality check on how your support strategy is landing overall.

what is a good NPS score?

It varies by industry, but generally:

• 0 or above = Solid
• 50+ = Excellent
• 70+ = World-class

how to improve NPS

• Close the loop with detractors. Acknowledge the feedback and take meaningful action—it can turn critics into loyalists.
• Turn promoters into amplifiers. Ask for reviews, referrals, or testimonials while the love is strong.
• Dig into the why. Pair NPS scores with verbatims and segment by channel, product line, or geography.

2. customer effort score (CES): the friction finder

what is CES?

Customer effort score (CES) is a metric that measures how easy it was for a customer to get their issue resolved or accomplish a task during an interaction with your company. Instead of asking how satisfied the customer was, CES focuses on the effort they had to expend.

It’s based on a simple, yet powerful question: “How easy was it to resolve your issue today?”

Customers respond on a scale, typically from 1 (very difficult) to 7 (very easy), though some use 1–5 or 1–10. The lower the effort, the better the experience.

why does it matter?

CES shines a light on one of the biggest barriers to customer loyalty: friction.

Research has found that making experiences effortless can have a significant impact on customer retention. Although other metrics like CSAT are also strong predictors of loyalty—and arguably more comprehensive—CES plays a crucial role by specifically targeting process simplicity.

While a high CES doesn't guarantee a high CSAT, a low CES almost always correlates with lower satisfaction. While customers might tolerate a robotic tone or a minor misstep, they are far less forgiving of confusing, time-consuming processes.

Bottom line: The more work your customers have to do to get help, the less likely they are to stick around.

how to improve your customer effort score:

• Remove dead ends: Review self-serve flows and ensure customers aren’t getting stuck in loops or unclear next steps.
• Empower your AI: Use automation to resolve common issues instantly, not just deflect or delay.
• Design for speed: Fewer clicks, fewer fields, and faster answers equal better outcomes.

3. average handle time (AHT): the misunderstood efficiency metric

What is AHT?

Average handle time measures the average amount of time it takes to handle a customer interaction—from the moment it starts to the moment it’s fully resolved. It includes:

• Talk time or chat time
• Hold time (if any)
• After-call work (like updating notes or following up)

AHT traditionally gave insight into the efficiency of support operations—but it’s not just about speed. The goal is to resolve customer issues clearly and completely, without unnecessary delays or rushed interactions.

Smart CX teams pair AHT with metrics like CSAT to get a full picture. Speed matters, but only when it's balanced with completeness, clarity, and care.

why does it matter?

At first glance, average handle time seems like the perfect efficiency metric. The shorter the interaction, the better... right? Not exactly.

In a traditional support model, where human agents were costly and queues were long, businesses often pushed to lower AHT at all costs. This sometimes led to rushed, transactional interactions that hurt the customer experience and drove up repeat contacts.

Today, with AI agents taking on more conversations, the dynamics are different:

• Speed still matters when a customer needs a quick, simple answer.
• Depth matters when a customer has more complex needs or wants a detailed conversation.

In some cases, a higher AHT could actually signal a great interaction: one where the customer is engaged, getting all their questions answered thoroughly, and building a stronger relationship with your brand.

Misusing AHT as a pure "speed race" metric still risks serious downsides:

• Incomplete resolutions: Rushing to close conversations leaves customers frustrated and forcing repeat contacts—which drives up costs and tanks CSAT.
• Poor customer experience: Fast but curt or confusing responses might check the AHT box, but they erode trust and loyalty over time.
• Missed opportunities for personalization: Rushing through interactions means missing signals for upsell, feedback, or deeper relationship-building.

AHT should be used as a diagnostic tool, not a target. A spike in AHT might mean a process is broken, a system is slow, or that your AI agent needs better training. AHT is not about being fast for the sake of speed. It’s about being efficient without sacrificing quality.

how to improve AHT (without tanking CSAT):

• Use AI to handle the repeatable stuff: Let automation take care of simple, high-volume topics so agents can focus on what requires expertise.
• Coach your AI agent for clarity: Rambling, vague, or indirect responses stretch out AHT. Clear, confident, relevant answers close tickets faster.

Design better flows: Route inquiries correctly, enable quick lookups, and remove redundant steps.

what your AI is trying to tell you (if you know how to listen)

Your AI agent is always sending signals. Every conversation it resolves, every moment it hesitates, every ticket it escalates—it’s data you can listen to, if you know where to look.

Metrics like CES, NPS, and AHT aren’t just numbers on a dashboard. They’re how your AI tells you what’s working, where it’s struggling, and what it needs to get better.

• Customer effort score (CES) shows when customers are gliding through a conversation — or silently struggling.
• Net promoter score (NPS) reveals whether your AI-led experiences are winning customer loyalty—or leaving them underwhelmed.
• Average handle time (AHT) surfaces how efficiently (and thoughtfully) your AI is resolving issues—not just how fast.

If you treat these metrics as conversation starters, not just scorecards, you’ll start to hear what your AI needs: better flows, smarter decision logic, faster escalations, deeper personalization.

Listen carefully, coach wisely—and you’ll build an AI agent that customers trust, prefer, and remember.

mastering AI customer service starts with mastering the right metrics

CES, NPS, and AHT aren’t just numbers on a dashboard—they’re your direct line into how your AI agent is performing, evolving, and impacting your customer relationships.

When you treat these metrics as coaching tools, not just performance indicators, you give your AI agent the ability to:

• Learn faster
• Resolve better
• Build deeper customer loyalty

It’s not just about measuring outcomes. It’s about making your AI better with every conversation. In short: master these metrics, and you’ll master your AI strategy.