QA as a service
"QA as a service" (QAaaS) covers three quite different things sold under one name. Knowing which one you're buying is most of the evaluation.
The three models
- Platform QAaaS — a testing platform you operate (often no-code, AI-assisted), sometimes with human testers on demand. You own the test library. Examples: Rainforest QA.
- Crowd QAaaS — a vendor orchestrates a large pool of testers on real devices in many countries. Breadth is the product. Examples: Testlio, Global App Testing, Applause, Test IO.
- Managed QAaaS — a vendor owns the outcome: either an automated suite they maintain (QA Wolf, MuukTest, Bug0) or, as we do, recorded human sessions with tracked coverage and evidence. See managed QA service.
What each costs
As of September 2, 2026, published starting prices in the managed tier run from about $2,500 to $5,000 per month; enterprise crowdtesting typically starts around $10,000; platforms price per run, credit, or seat. Almost everyone above a starter tier quotes rather than lists. Budget for the total — platform fee plus the people who operate it — not the sticker.
How to evaluate a QA-as-a-service vendor
- Who does the work, and can you name them? A crowd, a rotating bench, or one owner who learns your product.
- What is the evidence? Ask for a real finding. If there's no recording, screenshot, or exact state behind it, you're buying verdicts.
- How is coverage defined and reported? A percentage of automated flows, a device matrix, or a checklist with completion per session. Each answers a different question.
- Where do findings go? Into your tracker with reproduction steps, or into their portal.
- What about design quality? Accessibility, hierarchy, spacing, copy — most automation-first vendors don't report it at all.
- Can your coding agents use the output? If your team ships with Claude, GPT, or Gemini-based agents, findings should be structured for them. Ours are — see the format.
- What are the terms? Month to month vs. annual; who owns the artifacts; what you keep if you leave.
Where Agent Design Check fits
We are managed QAaaS of the human kind: a named QA owner runs recorded, checklist-driven sessions on your real product across desktop, tablet, mobile, and TV viewports, files findings with evidence into GitHub, Linear, Jira, Notion, Slack, or Teams, and reviews design quality along the way. No platform to run, no suite to inherit, month to month. It pairs well with an automated suite if you have one.