Most of us know Razer for its extravagant gaming peripherals with RGB lights and high-end keyboards, mice, headsets, and controllers. But it’s stepping into new territory in games with its AI tech. I checked out what was previously called Project Ava at CES 2025, an in-game AI assistant that works like a real-time guide or esports coach, but it was in a very early stage. While that has been officially re-revealed as Razer AI Gaming Copilot and is still in development, the project that could have a larger impact is the Razer AI QA Copilot.
It’s a program with an AI assistant built to streamline the quality assurance (QA) process in game development. The AI model is trained to detect and log bugs in real time as QA testers play the build of the game. It will also record clips of the moments it detects bugs so testers can see what happened and document log entries through video. This would let testers better determine if certain log entries are actually bugs or not and form their reports accordingly. You can watch the Razer AI QA Copilot in action in a short video teaser.
Razer said that it is intended to alleviate the more tedius aspects of the QA process and make it more efficient for testers and development teams. It’s not intended take the human element out of the process since AI simply cannot do the things an actual QA tester can do – like play the game itself, replicate the actions that trigger bugs, or evaluate the qualitative aspects of bug reporting.
Razer states that QA takes about 20% to 30% of development costs and one-third of development time is spent on QA, and QA Copilot is said to be able to identify 20% to 25% more bugs while accelerating test cycles by 30% to 50%. It also claims that AI Copilot can reduce a game’s time-to-market by 15% and bring cost-effective savings of 40% for teams. QA Copilot uses a form of spot detection, pinging its cloud servers at critical moments so it’s not constantly running in the background, which can theoretically keep energy consumption and total costs lower than they would otherwise.
The program is built to be a plug-in for existing game engines such as Unreal Engine, Unity, and custom engines through C++. It can track technical performance which includes frame rate, load times, and memory usage. There will also be genre templates so QA Copilot can better detect bugs for specific types of games.
Of course, development teams and games themselves are built differently so the actual impact would vary, also because Razer AI QA Copilot is still in early stages. The program will begin beta testing at the end of April for development teams, and Razer says it’s building the QA Copilot alongside devs to make it smarter and make sure it works as intended.
It’s worth remembering that the way AI models are used today may not be the way they are used in the future. And in light of a constant barrage of industry layoffs and jobs being lost, any use of AI should be met with a level of skepticism, especially in creative fields such as game development. Just take a look at the leaked Sony video of an AI model performing as the Horizon’s series’ lead character Aloy, and actor Ashly Burch speaking out about it. QA is an entirely different process of game development, but the human element that should not be lost.
