We are excited to announce that Scale is leading the $36M Series B of QA Wolf.
When it comes to testing software, there are two immemorial truths:
- Testing is absolutely crucial.
- No one wants to do it.
Tests can be the difference between catching a bug during CI/CD and shipping a bug that takes down your entire application. However, getting engineers to write and maintain tests can often feel like getting kids to eat their vegetables. They know they have to do it to get to dessert (shipping code) but it’s either done begrudgingly, or fed under the table to the dog (the summer intern).
End-to-end testing has a long history of being plagued by inefficiencies and headaches. Since the goal is to replicate end-user journeys, the most simple form is pure human labor. A dedicated team of people (either in house or outsourced) can click through the various expected flows of your application and report back any issues. The most obvious drawback of this approach is speed and scalability — your release process is inherently throttled waiting for manual feedback. As a result, companies have increasingly tried to turn to automating their end-to-end testing processes. Open source frameworks like Playwright, Cypress, and Selenium allow developers to write scripts that automate these clicks and actions to mimic a human user. The unfortunate reality, however, is that these automated tests are still hard to write, expensive to run, and notoriously flaky and brittle.
QA testing is a massive market. Thirty-five percent of IT budgets go towards quality and testing. Software investors don’t think about it much because the vast majority of this is labor, rather than tooling, and most of it is outsourced either overseas or to consulting firms. Behind any large software team is an even larger (albeit lower paid) QA team.
QA Wolf’s path to winning is AI-powered automation. We have a working thesis at Scale that the real winners in AI are yet to come, and that it takes a deep understanding of a problem to know how to most effectively deploy AI to meet needs. As Steve Jobs famously described, you have to start with the customer experience and work backward to the technology. QA Wolf has nailed the customer experience and is building incredible AI to support it. They’ve done this through a three-fold tech stack:
- Infinite backend: Tests are executed in a low-cost infinitely parallelizable cloud-native environment. Where most in-house test suites are executed in series, 100s of three minute tests running one after the other for hours, QA Wolf executes everything at the same time, surfacing issues in the three minutes it takes you to scan the HackerNews frontpage.
- Integrated frontend: QA Wolf engineers use a customer IDE with much more context specific to testing. Most treat QA as just another engineering task and reuse the same tools. But if a single engineer wants to process 10x the number of test failures there’s no time for downloading repos, syncing dependencies, and loading up an IDE for each one. Instead, the test plan, code, and failure context are streamed in real-time from the backend to a web-based app, allowing an engineer to process several test failures in parallel across a bunch of browser tabs.
- Intelligent automation: Most test failures aren’t touched by a human. QA Wolf has developed AI agents that replicate the same activities their human QA engineers do. Initially, these were the easy ones, like flaky tests and obvious errors, but what the team is learning is that in the world of AI, context is king. The same integrated experience that made the human QA engineers 10x productive is key to much higher rates of automated resolution.
The resulting customer experience is high test coverage, 24×7 response times, and zero flakes. The only thing your engineers have to deal with is fixing real bugs surfaced by the tests before they reach production.
It might sound like magic, and to customers, it certainly feels like it. Companies like Figma, Salesloft, and Drata (to name a few), have all managed to drastically improve test coverage and release time while cutting spend. For a market that’s best known for user frustration, their overwhelming customer satisfaction stands out.
