To evaluate a QA service provider, ask eight things before you sign: how the price is built, whether you can try them first, who starts and when, whether they work inside your sprints, whether the same people stay, how they prove AI, security and compliance testing, what happens if you leave, and whether they also build the product. A good provider answers each one with specifics. A vague answer is the red flag.
Most QA vendors look the same on a website: the same testing types, the same tools, the same certifications. They differ in eight places, and you only find them by asking. These are the questions that come up on almost every call we have with a buyer, with what a good answer sounds like and the answer that should end the conversation.
How is the price built, and what am I billed for each month?
A low hourly rate with unclear hours costs more than a higher rate you can check. You want to know what you pay for, not just how much.
A clear model (time and material, or a fixed team), a rate range by seniority, monthly invoices after the work, and hours you can see broken down by task.
A per-seat tool licence on top of the people, or “we’ll send a quote” with no range before a long discovery phase.
Can we try you before we commit?
A trial on your product tells you more than any case study. A vendor confident in its work makes trying cheap.
A short proof of concept on a slice of your product under NDA, or a first period you do not pay for if you are not happy, with agreed success criteria.
A minimum contract of several months before you see a single bug report.
How fast can you start, and with whom?
Your release does not wait for their hiring. Ask who would start, not just when.
A start date in weeks, named engineers you can meet, and honesty that contracts and access often take longer than finding the people.
“We’ll start recruiting once you sign.” You are paying for their hiring time.
Will your people work inside our sprints and tools?
Testing that happens after the sprint finds bugs after the sprint. The value is in being there when requirements are written.
They join your standups, your Jira or Linear, your Slack, and work on your project only. They read stories before code is written.
A separate portal you have to log into to see results, or engineers shared across several clients at once.
Who exactly will we get, and will they stay?
What a tester learns about your product is most of their value. Rotation throws it away every few months.
A straight answer about seniority (a senior works without being managed, a junior needs direction and costs less), and the same people for the life of the project.
A senior on the sales call and someone else on the project, or “the team” with no names.
Can you test our AI features, security and compliance, and prove it?
AI features fail in new ways, and regulated teams need evidence an auditor will accept, not just a pass rate.
Concrete examples (prompt injection, data leaking out, answers that sound right and are wrong), security testing as part of each release, and a per-release report of what was tested and what passed.
“Our AI tests everything.” Ask who checks what the AI produced. If the answer is nobody, keep looking.
What happens if we stop working with you?
The test suite is an asset you paid for. Lock-in shows up on the way out, so ask on the way in.
Automation lives in your repository from day one, test cases export to a standard format, and the notice period is weeks, not quarters.
Tests stored in the vendor’s own tool with no export, or a long notice period you only learn about in the contract.
Do you also fix the bugs you find?
A team that builds and tests the same code has a reason to look away. Independence is what makes a bug report trustworthy.
No: they test, and each report comes with steps, evidence and the likely cause so your developers fix it fast.
“We can do development too.” Then ask who reviews their code.
- Price model, rate range by seniority, monthly invoice after the work
- A trial or proof of concept with success criteria
- A start date in weeks and named engineers
- Inside your sprints and tools, on your project only
- Seniority stated plainly, same people for the whole project
- Real examples for AI, security and compliance, and a per-release report
- Tests in your repository, standard export, notice in weeks
- They test, they do not build your product
If you want to put us through these eight questions, book a 30-minute call.