QA software pricing is usually a subscription billed per seat, per reviewer or per conversation scored, with plan tiers that add AI features, automation limits, integrations and support. What you pay depends on who counts as a seat, how many conversations you handle on which channels, how much of the scoring is automated and what the contract includes.
In short
- Four pricing models cover most quality assurance tools: per seat, per reviewer, usage-based and a platform fee with usage on top.
- The definition of a seat moves the bill more than the price per seat, so get it in writing.
- The main cost drivers are team size, conversation volume and channels, AI scope, integrations, data terms, support and contract length.
- A worked example shows how the same 560-person team is counted under each model, and a break-even rule tells you when usage pricing beats per-seat pricing.
- The license is not the whole cost: reviewer time, onboarding and rubric work belong in the comparison too.
If you are here for our own numbers, the plans are on the Kaizo pricing page, and the section on Kaizo below explains how they work. The rest of this guide applies to any vendor.
How QA software pricing works: the four models
QA software is priced in one of four ways: per seat, per reviewer, per unit of usage, or a platform fee plus usage. Each counts something different, so the same team can look cheap under one model and expensive under another.
| Model | What is counted | Scales with | Fits best when | Watch out for |
|---|---|---|---|---|
| Per seat | Every user who logs in: agents, reviewers, managers | Headcount | Headcount is stable and every agent sees their scores | Seasonal agents and read-only users counted as full seats |
| Per reviewer | Only the people who score | QA team size | Manual review, a small QA team, agents who never log in | Price rises fast when you add automated scoring or open access to agents |
| Usage-based | Conversations scored, or minutes of audio processed | Volume | Volume is low or seasonal relative to headcount | Bills that grow with every new channel or AI agent conversation |
| Platform fee plus usage | A fixed fee, then volume above an allowance | Both | You want a predictable base and growth room | Overage rates and allowances that reset monthly |
Most tools sell these in tiers. The higher tier usually adds more automated criteria, deeper analytics, integrations, data controls and a named account contact, rather than a different pricing unit.
What counts as a seat
A seat is whoever the contract says needs a license, and vendors draw that line differently. Some count only reviewers, some count every agent being scored, and some also count team leads and managers who view dashboards.
Ask these questions before you compare any per-seat price:
- Does an agent who is scored but never logs in count as a seat?
- Do team leads and managers who only read dashboards count?
- How are seasonal or temporary agents counted: at peak, on average or per month?
- In a BPO, are each client’s agents counted separately?
- Does an AI agent count as a seat, or are its conversations billed as usage?
Write your own org chart in the vendor’s terms. A team of 500 agents with 20 reviewers and 40 team leads is 20 seats under one definition and 560 under another.
What drives the cost of QA software
The cost of QA software is driven by eight things, led by who uses it, how many conversations it scores, on which channels and how much of the scoring AI does.
| Driver | Why it moves the price | What to have ready |
|---|---|---|
| Team size and roles | Seats or reviewers are the main unit in most models | Headcount by role, peak and average |
| Conversation volume | Usage models bill it directly; seat models price it into tiers | Monthly conversations per channel |
| Channels, especially voice | Calls need recording import and transcription | Call minutes per month and languages |
| AI scope | Fully automated scoring of every conversation does more work than AI that assists a reviewer | The share of conversations you want scored, and how |
| Automated criteria | Plans often cap how many criteria the AI scores | Your scorecard, with the criteria you want automated |
| Integrations | Native helpdesk integrations differ from custom connectors and exports | Your helpdesk, in-house systems and BI tools |
| Data terms | Custom retention, storage location and audit logs are often top-tier features | Your security and data residency requirements |
| Support and contract | A named success manager, business reviews, SLAs, term length | The support level you need, and the term you can commit to |
Volume and AI scope are where buyers most often under-specify. A quote for a tool that scores a sample is not comparable with a quote for one that scores every conversation, even at the same seat price.
A worked example: one team, four pricing models
This is an illustrative example with a made-up team, and it uses no vendor’s prices. It shows how the unit you are billed on changes with the model, and how to find the point where one model becomes cheaper than another.
The team: 500 agents, 20 QA reviewers and 40 team leads, handling 120,000 conversations a month across chat, email and calls. An AI agent handles another 30,000 conversations a month.
| Model | Units billed per month | What happens with 100 extra seasonal agents for 3 months | What happens when the AI agent’s volume doubles |
|---|---|---|---|
| Per seat (all users) | 560 seats | 660 seats during peak, if counted monthly | No change, unless AI agents are billed separately |
| Per reviewer | 20 reviewers | No change, unless you add reviewers | No change, but someone has to review the extra volume |
| Usage-based (all conversations) | 150,000 conversations | Rises with the extra contacts those agents handle | 180,000 conversations |
| Platform fee plus usage | Fee plus volume above the allowance | Overage if peak volume exceeds the allowance | Overage on 30,000 more conversations |
The break-even rule between per-seat and usage pricing
Per-seat and usage pricing cost the same when the seat price times the seats equals the per-conversation price times the conversations. So the break-even price per conversation is the monthly seat price times seats, divided by monthly conversations.
For this team that is 560 seats ÷ 150,000 conversations = 0.0037. In plain terms: a usage price below about 1/268 of the monthly seat price is cheaper than per-seat pricing for this team, and above it, per seat wins. A team with fewer conversations per person flips the result.
The cost a license price does not show
Reviewer time is often the bigger number. Say the team above reviews a 3% sample of its human-handled conversations by hand, the sample size we examine in our post on why a 3% QA sample cannot support agent-level decisions.
That is 3,600 reviews a month, and at an assumed 12 minutes each it is 720 reviewer hours a month for a sample, which is why the coverage you buy matters as much as the price.
Our manual QA vs Auto QA comparison has a cost model you can run with your own numbers.
Costs outside the license
The total cost of QA software includes onboarding, rubric work, integration, reviewer time and the time to reach full coverage. A cheaper license that needs more manual review can cost more over the contract.
| Cost | Who carries it | How to reduce it |
|---|---|---|
| Onboarding and setup | Vendor and your QA lead | Ask what onboarding includes on your plan |
| Building or migrating the scorecard | Your QA team, with the vendor’s help | Bring your current scorecard; see our scorecard feature page for how criteria carry over |
| Testing the AI against your reviewers | Your QA team | A short validation plan before go-live |
| Integration work | Your IT team, or the vendor for custom connectors | Prefer native integrations for your helpdesk |
| Reviewer time after go-live | Your QA team | Automate the routine checks, keep people on judgement calls |
| Change management | Team leads | A rollout plan and early communication to agents |
For the value side of the same calculation, use our guide to the ROI of Auto QA, which turns these costs into a payback model for finance.
Questions to ask before you get a quote
Ask every vendor the same seven questions so their quotes can be compared on one basis. Send them with your headcount and volume, and ask for the answer in writing.
- What is the pricing unit, and who counts as a seat?
- What share of conversations does each plan score, and how is each scored?
- How many automated criteria does each plan include?
- Which channels and integrations are included, and which cost extra?
- Are AI agent conversations included, and how are they billed?
- What does onboarding include, and is there a setup fee?
- What changes at renewal, or when we add seats or channels mid-term?
If you are comparing named tools at the same time, our comparison of customer service QA software covers what each one does well.
How Kaizo pricing works
Kaizo bills per seat, and every plan scores 100% of your conversations, as set out on the Kaizo plans page. The plans differ in how much analysis, coaching, automation and customization come with it, and the exact price comes as a quote based on your team size.
| Plan | Seats | What it adds |
|---|---|---|
| Insight | 1 to 25 | AutoPilot QA on every conversation, dashboards, AI summaries and on-demand transcription, coaching, a ready-to-use scorecard, 10 custom automated criteria, assignments and calibration |
| Intelligence | 26 and more | Everything in Insight, plus AI coaching cards, 25 custom automated criteria, automated BI exports, extended AI insights, a dedicated success and account manager and annual business reviews |
| Enterprise | Unlimited | Everything in Intelligence, plus third-party integrations, custom dashboards, custom data retention and storage location, 50 custom automated criteria, white-glove onboarding and AI workshops, enterprise support SLAs and audit logs |
A seat on Kaizo is anyone who needs to log in: agents being reviewed, reviewers and managers. Every plan includes help building your rubric and connecting your helpdesk. There is no self-serve trial; instead, we show the product in a live demo with realistic sample conversations.
Why no price on the page? Seat count and ticket volume change the number enough that a headline price would mislead you. Teams commonly start on Insight and move to Intelligence once QA runs unattended and they want coaching and exports.
How to get an accurate QA software quote
To get an accurate quote, send each vendor your headcount by role, monthly conversations per channel, your helpdesk, the share of conversations you want scored and your data requirements. With those five facts, a vendor can price your team rather than a typical one.
Get an exact number for your team. Tell us your team size and helpdesk, and we will come back with a quote and a live demo. Let’s talk.
Frequently asked questions
How much does QA software cost?
It depends on the pricing unit, your team size, conversation volume, channels and how much of the scoring is automated. Many vendors quote larger teams rather than publish a single price, because seat definitions and volume change the total. Compare quotes on the same headcount, volume and coverage, and add reviewer time to the license cost.
Is per-seat or usage-based pricing better for QA software?
Per-seat pricing is more predictable when headcount is stable and most people log in. Usage-based pricing can be cheaper when volume is low relative to headcount, and dearer when volume grows or AI agents add conversations. Use the break-even rule: monthly seat price times seats, divided by monthly conversations.
Do QA tools charge for AI agent conversations?
Some do and some do not, and the answer often sits in the definition of a seat or a billable conversation. Ask whether AI agent conversations are scored on the same plan, whether they count as usage, and what happens to the price if your bot’s share of volume grows during the contract.
Is there a free trial for QA software?
Some QA tools offer self-serve trials and others, including Kaizo, run a live demo instead. A demo on realistic sample data shows the full product with scorecards and dashboards already populated, which an empty trial account rarely does. After purchase, plan a short validation of the AI against your own reviewers.
Can you change QA software plans later?
Usually yes, but check the terms for moving up or down during a contract and at renewal. On Kaizo, teams commonly start on Insight and move to Intelligence once QA runs unattended and they want coaching cards and automated exports. Ask any vendor how mid-term seat growth is priced.
