Greet the customer. Verify identity. Listen. Solve the problem. Close cleanly. Every QA checklist in circulation says roughly this, and none of them are wrong.
The difference between a QA program that changes behavior and one that generates paperwork was never the criteria. It is whether those criteria reach every call and mean the same thing every time. Most operations score one or two calls per agent per week, assigned to whoever had capacity, with each reviewer reading "showed empathy" a little differently. Scored that way, a checklist describes a sample. Scored the same way on every call, it becomes the standard the floor actually runs on.
What is a call center quality assurance checklist?
A call center quality assurance checklist is the list of criteria an evaluator checks when reviewing a customer interaction: did the agent verify the caller, deliver the required disclosures, diagnose the actual issue, resolve it, and close it out correctly. It walks the call in order, from opening to after-call work, so that evaluations stay consistent from reviewer to reviewer and from call to call.
There are two kinds, and most published templates only cover the first. The evaluative checklist is used after the call, by QA, to score what happened. The preventive checklist is used by the agent, before and during the call, to get the interaction right the first time. This guide includes both.
The call center QA checklist: what to evaluate
Use the categories below as written or trim them per queue. One rule governs every item: it must be answerable yes or no from the recording. "Showed empathy" fails that test; "acknowledged the customer's frustration before starting the fix" passes it.
1. Opening and verification
- Greeting — states their name and the company in the first line
- Verification — completes every required identity check before any account detail is discussed
- Reason capture — restates why the customer is calling before starting to solve it
2. Discovery and diagnosis
- Probing — asks questions that test the stated problem instead of accepting it at face value
- Actual cause — identifies what is actually wrong, which is often different from what the customer first describes
- Confirmation — restates the diagnosis and gets the customer's agreement before the fix begins
3. Communication
- Plain language — explains without internal system names or process jargon
- Listening — does not talk over the customer or ask for information the customer already gave
- Hold and transfer discipline — asks before placing a hold, sets a time expectation, and transfers with context so the customer never repeats the story
4. Resolution and accuracy
- Correct answer — the solution matches policy and actually resolves the issue
- Complete record — every system that needed updating was updated during or immediately after the call
- Expectations — the customer leaves knowing exactly what happens next and when
5. Compliance
- Disclosures — required statements delivered fully, word for word where policy demands it
- Boundaries — no guarantees, legal opinions, or financial advice outside policy
- Data handling — no sensitive data read back aloud or left in notes where it does not belong
6. Closing and after-call work
- Resolution check — asks whether the issue is fully handled before ending the call
- Further needs — offers additional help once, without scripted over-asking
- Wrap-up — disposition and notes are accurate and finished within the expected window
Eighteen items across six categories is enough to cover compliance and resolution without turning each evaluation into an afternoon.
The agent-side checklist: before the call
The evaluative checklist grades the call after it happens. The preventive version runs before it, and it belongs to the agent:
- Systems ready — every tool the queue requires is open and logged in before the first call
- Last feedback in view — the one behavior flagged in the most recent coaching note is written down where the agent can see it
- Queue context — the agent knows today's known issues (an outage, a promotion, a billing-cycle date) that will be driving calls
One more decision matters here: the evaluation criteria should not be a secret. Agents who can see the checklist treat it as the structure of a good call rather than a trap waiting in next week's review. The fastest way to make the checklist improve calls is to hand it to the people taking them, and to let new agents rehearse against it in simulation before their first live call.
Call center audit checklist: is the QA program itself working?
The checklist evaluates calls. Once a quarter, something has to evaluate the checklist. Five questions:
- Coverage — what percentage of last month's interactions were actually evaluated? At single-digit coverage the program is a sample audit, whatever it is called internally.
- Consistency — the last time two evaluators scored the same call, how far apart were they?
- Freshness — how many days pass between a call and the agent seeing feedback on it?
- Drift — when a policy changed, did the checklist change the same week, or is it still scoring against the old process?
- Downstream use — can you name a coaching action from last month that traces to a specific checklist item?
If two or more of these produce uncomfortable answers, fixing the program will move quality more than rewriting the items.
How to apply the checklist consistently
Three failure modes account for most checklists that exist but change nothing.
The first is sampling. When one or two calls per agent per week get evaluated, an agent's score reflects the luck of the draw, and everyone on the floor knows it. The percentage of calls evaluated is the credibility of the program.
The second is evaluator variance. Hand the same call to three reviewers and you will often get three different scores, because each reads "professional tone" through their own ears. The fix is twofold: word every item as an observable behavior, and calibrate reviewers against the same calls until their scores converge. A checklist without calibration is a collection of opinions in table form.
The third is latency. Feedback that arrives two weeks after the call describes a conversation the agent no longer remembers. It gets read, filed, and repeated.
This is the case for automating the application rather than the writing. Reddy's Auto QA applies your checklist to every interaction, calibrated against your own evaluators' scoring, and routes what it finds into coaching while the call is still fresh. The checklist stays yours. What changes is that it finally gets applied the way it was written: to everything, identically.
Checklist vs. scorecard
The two get used interchangeably and they are not the same instrument. The checklist is the list of what to check, walked in call order. The scorecard is the scoring layer on top: weights, point values, pass/fail rules, and what an overall number means. Draft the checklist first. Formalize weights once your evaluators agree on what the items mean, because weighting criteria nobody reads the same way just produces precise disagreement.
Frequently Asked Questions
Run it on every call
A checklist applied to two percent of calls describes two percent of your operation. Reddy applies yours to all of it, calibrated to your evaluators, with feedback that reaches agents while the call still matters.