Every contact center leader I know has had this conversation with finance. "We want to monitor more calls." "Great. How many more reviewers do you need?"
For a long time, that was the only answer. More calls monitored meant more people listening. Headcount grew with call volume, and QA was always the first budget to get squeezed.
That's not the only answer anymore. This post covers how to think about call monitoring at scale: sampling versus broad coverage, what to score for compliance versus what to score for coaching, and how your reviewers' time shifts when software does most of the listening.
The old math of call monitoring
Manual call monitoring is simple. A reviewer pulls a recording, listens to it, scores it against the form, and writes notes. Then they pull the next one.
Each call takes at least as long as the call itself. Add scoring and notes on top. A reviewer can only get through so many calls in a shift, no matter how good they are.
So most teams sample. A handful of calls per agent per month. That's not laziness. It's the only thing the math allows.
Sampling: what it gets right and wrong
I want to be fair to sampling, because a good sample is useful.
What sampling gets right:
- It's cheap to start. You need a form, a reviewer and a recording system.
- It can tell you about the team. If you pull a random sample across the floor, you get a decent read on overall quality trends.
- Reviewers can go deep. With fewer calls to cover, each review can be thorough.
I've written before about how many calls QA should score if you're building a sample. The short version: be deliberate about what you pull and why.
Where sampling falls short:
- It's thin per agent. A few calls a month can't tell you much about one person. One bad call swings the whole score.
- It misses rare events. A compliance miss that happens on a small share of calls may never land in your sample. The calls that matter most legally are often the ones you don't pull.
- It's slow. Sampled calls can be weeks old by the time they get coached.
- Agents feel singled out. "Why did you pick that call?" is a fair question, and a hard one to answer when the pick feels random.
Broad coverage: what changes
Call monitoring software that scores automatically changes the math. The software does the listening. Your reviewers don't scale with call volume anymore.
When you can score far more of your calls, a few things happen:
- Agent scores get more stable. A score built on many calls is harder to argue with than a score built on four.
- Rare problems surface. If a compliance step is getting skipped, you're much more likely to catch it.
- Feedback gets faster. Supervisors can coach on recent calls.
- You see patterns. Which question fails most often? Which call type trips people up? Which shift has the problems?
Broad coverage isn't automatically better, though. If your form is vague, scoring more calls just gives you more vague scores. Fix the form first. Then scale it.
What to score for compliance vs coaching
This is where a lot of teams get it wrong. They treat compliance and coaching as one thing on one form. They're different jobs.
Compliance items
Compliance questions are usually yes or no. Did the agent read the required disclosure? Did they verify identity before discussing the account? Did they avoid the words they're not allowed to say?
These are the items where coverage matters most:
- They're objective. It happened or it didn't. That makes them a good fit for automated scoring.
- The risk is concentrated. One miss can cost real money or a client relationship. Broad coverage is how you catch them.
- They should be flagged fast. A missed disclosure should reach a supervisor quickly, not show up in a monthly report.
Keep your compliance list short and specific. If you can't write a question so two people would score it the same way, it's not a compliance item yet.
Coaching items
Coaching questions are about how the call went. Did the agent show empathy? Did they control the call? Did they explain the next steps clearly? Did they try to resolve it on the first contact?
These items work differently:
- They're about trends, not single calls. One so-so greeting doesn't matter. A pattern of rushed greetings does.
- They need context. A short, direct call might be exactly what the customer wanted.
- They feed conversations. The goal isn't a number. The goal is a supervisor sitting with an agent and working on one thing.
With broad coverage, coaching items get more useful because you're looking at a real pattern instead of guessing from a handful of calls. A supervisor can say, "You do great on the open, but you're skipping the recap on billing calls. Here are three examples." That lands differently than "your score dropped."
Keep them separate in reporting
Score both on the same call if you want, but report them separately. A compliance miss is an escalation. A low coaching score is a conversation. Mixing them into one number hides both.
How reviewers' time shifts
This is the question QA teams care about most. If software does the listening, what do the reviewers do?
Here's how I see the work changing.
Before: most of the day is playback. Pull a call, listen, score, write notes, repeat.
After: the day looks more like this.
- Calibration. Reviewers check that the software reads the form the way your team does. They compare its scores to their own on a set of calls and tighten question wording where needed. More on that in how to calibrate AI scoring.
- Reviewing flagged calls. Compliance misses, low scores and odd calls get routed to a person. The reviewer listens to the calls that matter instead of random ones.
- Handling disputes. When an agent pushes back on a score, a person reviews it and decides.
- Working with supervisors. Reviewers turn patterns into coaching plans. "Here's what's failing on this team, and here are the calls that show it."
- Form maintenance. Scripts change, products change, regulations change. Somebody has to keep the form current.
That's better work. It uses what your best reviewers actually know, instead of burning their day on playback.
It also means you don't need to grow the QA team every time call volume grows. The same reviewers can support a much bigger operation because their time goes to judgment, not listening.
A simple way to start
You don't have to switch everything at once.
- Clean up your form. Separate compliance items from coaching items. Rewrite anything vague.
- Pick your calibration calls. Pull 10 to 20 calls your team already scored. Mix good, average and poor calls.
- Calibrate. Compare software scores to your team's scores line by line. Fix the gaps.
- Scale up. Expand coverage once the scores line up.
- Shift reviewer time. Move reviewers from playback toward flagged calls, disputes and coaching support.
Where Otto Fits
OttoQA scores your calls against your own QA form, in any language, at 96% accuracy against your team's scoring. That lets you monitor far more calls without adding reviewers. We calibrate using 10 to 20 calls your team already scored.
No IT needed. It works with any CCaaS or cloud CRM via API, including Zoom Contact Center. Setup takes about 2 weeks. Hosting is US-only, and your customers' calls are never used for training. We're PCI compliant and sign HIPAA BAAs, which matters when you're monitoring calls with payment or health information on them.
The one-time $1,000 covers setup and a 200 to 500 call proof of concept. After that, it's contract or month-to-month. Details are on our pricing page.
If you want to see what broader coverage looks like on your own calls, Book a demo.
Frequently asked questions
What is call monitoring at scale?
It means reviewing a large share of your calls instead of a small sample, usually with call monitoring software that scores calls automatically so reviewers don't have to listen to them all.
Is sampling calls still useful?
Yes. A deliberate sample can show team-level trends. It's thin for judging individual agents and can miss rare compliance problems.
What should I score for compliance versus coaching?
Compliance items are objective yes-or-no checks, like required disclosures and identity verification. Coaching items are about how the call went, like empathy, call control and clear next steps. Report them separately.
Will call monitoring software replace my QA reviewers?
No. Their work shifts from listening to calibration, reviewing flagged calls, handling disputes and helping supervisors coach.
How long does it take to set up call monitoring software like OttoQA?
Setup takes about 2 weeks, including calibration on 10 to 20 calls your team already scored.
About the author
Tom Laird is the founder of OttoQA and CEO of Expivia, with about 30 years in contact centers. He hosts the Advice from a Call Center Geek podcast. Connect with Tom on LinkedIn.