You can tell a lot about a contact centre within five minutes.
The dashboard won’t show you. The roadmap won’t either. But the way the place sounds will. You can hear whether a team is firing on all cylinders or just treading water. It’s in the energy of the floor. The way agents talk about tricky calls. The questions team leaders ask each other in passing.
Almost everywhere, the ambition is sky-high. Leaders talk about improved journeys, better tools, lower pressure, higher quality. Nearly 90 percent of organisations say they’re using AI somewhere. Yet only a quarter rank it as their number-one priority.
Spend half an hour inside though, and the cracks show up.
The tools may have moved on, but the way people work hasn’t changed much. Agents still hunt for answers through three different systems. Decisions get stuck in approval loops. Knowledge lives in several places at once, and nobody’s quite sure which version is current.
The pressure builds. People improvise. The cycle repeats.
Beneath all of this sit five human skills that shape every good service moment: Listen, Think, Speak, Do and Improve.
If you think about it, they seem simple. Yet they decide whether AI helps or just makes things more difficult.
These skills are the operational backbone of the contact centre. And most organisations are still struggling to get them right.
Here’s why each one matters, where teams trip up and what changes when you get them right across the three moments that shape every interaction: before the customer calls, during the conversation, and after they hang up.
1. Listen: hearing what customers need (before they finish explaining)
Listening means understanding what customers need and what’s making them tense. Often before they’ve finished their second sentence.
Most centres figure out what’s really wrong halfway through the conversation. By then, customers have repeated themselves twice. The clues were there before the phone rang (weekend bookings failing, product issues creeping in), but people were too busy putting out fires.
During calls, agents rely on gut feel because the system gives them nothing. After calls, the data arrives three weeks late in reports nobody reads.
A local council had a spike in calls every Monday morning. The explanation: “People always call after the weekend.” Eventually someone checked what residents were calling about. A specific bin collection round kept getting missed on Fridays. Residents waited all weekend, then called Monday morning. Fix the missed collections and the spike disappeared.
AI can spot these patterns in real-time, surfacing what’s driving contacts before agents even pick up the phone, but only if you’re set up to catch the signals.
2. Think: making decisions with confidence (not guesswork)
Thinking means making the right call quickly. The path is clear, documented and supported. When people know where an issue should go, decisions start feeling routine.
The “right way” to handle a case often lives in one person’s head. Without it written down, every tricky call becomes guesswork. Agents get stuck choosing between options, none quite right. They transfer, escalate, or put people on hold while they ask someone else.
A renewals team discovered two agents giving completely different answers to identical questions. Neither was wrong, they’d just learned different versions from different people. Once they agreed on one path, handle times dropped and conversions jumped. Nobody worked harder, they just stopped second-guessing themselves.
AI can route cases to the right place and suggest the next best action, but only if you’ve documented what “right” looks like.
3. Speak: saying the right thing at the right moment
Speaking is where customers feel your brand. Where being helpful, clear and doing the right thing either come together or fall apart.
Your knowledge base hasn’t been touched since 2022. Scripts sound stiff. Letting agents make it up is risky. So agents wing it, say sorry, change their answer, correct themselves. Customers can tell they’re unsure.
A bank found customers kept calling back to double-check their mortgage terms. The information was correct. The phrasing made people nervous. One small change to how agents explained the numbers (same facts, clearer words) and the callbacks stopped.
AI can suggest the right words in the moment, but only if your knowledge is current and your tone is defined
4. Do: getting things done without the scramble
Doing is where the day is most likely to fall apart. Resolving issues, updating systems, filling forms, chasing follow-ups. Even the best agents struggle if the workflow is held together with sticky notes and hope.
Workflows aren’t mapped. Systems don’t talk. Every agent does it differently. During calls, agents toggle between screens, copy-paste, get lost in forms. Loose ends pile up. Follow-ups get forgotten.
An insurance team spent months trying to speed up fraud checks. Someone drew the whole process on a whiteboard. Half the team checked credit first, then verified identity. The other half did it backwards. Same result, but one route took three extra minutes and hit two extra systems.
AI can automate the workflow, but only if there’s a clear, standard process to automate.
5. Improve: making tomorrow easier than today (not just talking about it)
Improvement decides whether you’re learning or just repeating the same problems.
Nobody knows how busy it’s going to be. Staffing is guesswork. Things go sideways and nobody notices until it’s too late. Coaching waits until someone has time. Quality checks a handful of calls. Feedback arrives too late. Issues repeat like a broken record.
The contact centres that break this cycle spot patterns fast, update knowledge while it’s warm, and coach from every call. They make fixing things part of the job, not a special project.
A utilities company dropped handling time by 15 percent after fixing one broken knowledge article. It had been wrong for months but nobody touched it because “that’s how we’ve always done it.” One edit, thousands of minutes saved.
AI can flag what needs fixing as it happens, but only if you’re ready to act on it quickly.
Making it work
These five skills play out across three moments: before contact, during contact, after contact. Get them right and AI will change your business for the better. Get them wrong and AI is just another failed pilot.
Right now, 95% of pilots fail to show a positive impact. Most of this comes down to readiness. Which means there’s a real opportunity to join the 5% that actually work.
Start by looking at five everyday examples from your operation. Where do agents get stuck? Where do customers have to repeat themselves? Where does a fix that worked Monday stop working by Friday?
Tackle things in order: quick wins first (better knowledge, clean summaries), structural stuff second (better routing, real-time support), big moves last (automation that scales).
Build a flow. Short coaching sessions. Quick knowledge amends. Small improvements delivered often. When fixing things becomes part of the week instead of a project, the five skills get stronger naturally.
Get a clearer view of where you stand
Cirrus runs a one-hour session that shows how your operation performs across these five skills and three moments.
You’ll see where the work is solid, where it’s slipping and what to tackle first to make tomorrow easier than today.
No perfect data needed. No endless prep.
Just a straight look at how your day works and what needs improving before you add more tech to the pile.
