Skip to content
DIRI AI
All guides

Guide · Conversation practice

Call Centre KPIs (AHT, FCR) and How Training Affects Them

What AHT, FCR, CSAT and other call centre KPIs measure, how phone skills affect each one, and how to link practice to metrics without gaming them.

Call centres run on numbers. Average handle time, first contact resolution, customer satisfaction and quality scores appear on dashboards every morning. The numbers are useful, but they are easy to misread, and they are easy to push in the wrong direction. This guide explains the main KPIs, what behaviours sit underneath them and how training and practice can influence them in a healthy way.

Definitions vary between organisations, so check how your own company calculates each one.

The main KPIs in plain language

Average handle time (AHT). The average length of a customer interaction, usually including talk time, hold time and the work done after the call. It is a measure of efficiency, not quality.

First contact resolution (FCR). The share of customer issues solved in the first contact, without a repeat call or follow-up. Often measured by checking whether the same customer contacts you again about the same issue within a set period.

Customer satisfaction (CSAT). A rating given by the customer, usually through a short survey after the call.

Quality or QA score. An internal rating from reviewers who score calls against a checklist. A handful of clear criteria works better than a long checklist.

Service level and abandonment. How quickly calls are answered and how many callers give up. These are mostly driven by staffing, not agent skill.

Occupancy and adherence. How much of an agent's time is spent on calls and how closely they follow their schedule.

The trap of single metrics

Every KPI invites shortcuts when it becomes the only target.

  • Push AHT alone and agents may rush callers, which hurts FCR and satisfaction.
  • Push FCR alone and agents may stay on long calls even when escalating would help.
  • Push CSAT alone and agents may over-promise.

Healthy management uses a small set of metrics together and looks at the relationships between them. If AHT drops but repeat calls rise, you have not become more efficient. You have moved the work.

Behaviours that sit under the numbers

Training works on behaviours, so it helps to link each KPI to the habits that influence it.

For AHT

  • Opening cleanly without unnecessary small talk
  • Asking focused questions in a sensible order
  • Knowing the systems and where to find information
  • Handling holds and transfers smoothly
  • Closing with a clear recap instead of drifting

For FCR

  • Clarifying the real issue before offering a fix
  • Confirming the customer's understanding
  • Checking for related problems before ending the call
  • Knowing when and how to escalate properly
  • Writing notes the next agent can use

For CSAT

  • A warm and steady tone
  • Listening without interrupting
  • Showing the customer what happens next
  • Keeping promises on follow-up
  • Handling frustration calmly

For QA scores

  • Following required steps and statements
  • Accuracy of information
  • Compliance with policy

How training moves each metric

Be careful with promises here. Training does not guarantee numbers, and results depend on your processes, tools and staffing as much as agent skill. What you can say is that practising specific behaviours makes them easier to do consistently, which gives the metrics a chance to improve.

Link one behaviour to one metric. Rather than a generic "improve quality" workshop, pick a pattern from your data. If repeat calls are high, practise confirming understanding and recapping the next steps.

Practise the moments, not the whole call. A two-minute drill on a transfer script gives more benefit than rereading a policy document. Our guide on customer service phone training outlines how to build short, regular drills.

Use realistic customers. Agents behave differently under mild pressure. Practice with a difficult customer, such as someone confused about an invoice or irritated about a delay, builds composure.

Let agents practise privately. Some people avoid role play with colleagues. In DIRI AI, the AI plays the customer on a voice call and the agent receives a report with specific moments to retry. The sample dialogues show an example, and how scoring works explains the criteria.

Review results with context. After a few weeks, compare the targeted behaviour and the related metric. Keep in mind that other changes, like seasonality or new products, may also move the numbers.

Building a simple measurement plan

  1. Pick one metric and one behaviour. For example, FCR and confirming understanding.
  2. Set a baseline. Note where you are before the change.
  3. Run short practice for a few weeks. Five to ten minutes, several times a week.
  4. Sample real calls. Check whether the behaviour shows up.
  5. Review the metric, and the metrics next to it, to catch side effects.
  6. Adjust. Keep what helps, drop what does not.

Avoid declaring success from a single week of data. Metrics move for many reasons.

Using KPIs fairly with agents

  • Explain what each metric means and why it matters.
  • Share trends rather than ranking people publicly.
  • Pair numbers with coaching, not just pressure.
  • Recognise improvement, not only top scores.
  • Remember that complex calls naturally take longer.

Agents who understand the purpose behind a number are less likely to game it. For ideas on a structured onboarding approach, see the guide on practising customer service calls.

Related: Call Centre QA Scorecard Template (Sales + Service), AI training for call center agents, Customer Service Phone Training. Free tool: Sales Glossary.

FAQ

What is a good AHT?

It depends on your industry and the kind of calls you handle. Compare against your own history and call types rather than generic benchmarks.

Can training reduce handle time without hurting quality?

It can help by making routine steps smoother. Watch FCR and satisfaction alongside AHT to make sure quality holds.

How long until training shows up in the metrics?

Usually several weeks, and the effect is hard to separate from other changes. Track the behaviour as well as the metric.

Should every agent have the same targets?

Not necessarily. New hires, specialists and complex queues often need different expectations.

To give your team a place to rehearse openings, transfers and difficult customers, try DIRI AI. The Free plan gives you 3 voice calls a month of up to five minutes each, with no card needed. Start free and choose a customer service scenario.