Side, a seasoned services provider, has entered into a six‑month memorandum of understanding with Denmark‑based artificial‑intelligence specialist Modl.ai. The agreement paves the way for a joint effort to build an AI‑driven quality‑assurance (QA) platform that blends Side’s extensive testing expertise with Modl.ai’s cutting‑edge machine‑learning capabilities. Side, previously known as PTW, has been operating in the outsourcing and testing arena for many years, delivering manual and automated QA services to game developers worldwide. Modl.ai, by contrast, is a relatively new player in the gaming sector.
Founded in 2018, the startup secured $1.7 million in seed funding in 2020 and has since been experimenting with AI‑assisted testing tools. "The MOU essentially combines the strongest assets of both companies into a singular, differentiated solution," explains Martin McBride, Side’s chief commercial officer.
"For the games industry, one of the holy‑grail ambitions has been to harness AI to augment QA in a meaningful way. Nobody has cracked that problem yet, and we believe we are on the brink of doing so." Earlier AI‑enabled QA approaches—some of which were also pioneered by Modl.ai—required developers to ship full builds or even source code to the testing vendor. Those solutions relied on deep integration via plugins and SDKs, creating friction and, as founder Christoffer Holmgard notes, becoming increasingly unwieldy over time.
Holmgard emphasizes that the new system eliminates the need for any internal access. "We treat the product exactly as you would hand it to a QA tester or eventually to a consumer.
The game runs on a comparable device, our AI watches the screen, simulates key inputs, records everything that happens, and then passes that data to a second AI layer that flags potential bugs. That layer then escalates items to a human reviewer for final validation." The first phase of the workflow is what Modl.ai calls "building game knowledge." Developers upload assets, design documents, screenshots, or even short video clips and converse with the AI in natural language. The system absorbs this information, forming a contextual model of the game’s mechanics, art style, and narrative.
"All interaction is in plain English," Holmgard explains. "We chat with the AI about the game, feed it images, feed it design docs—anything that helps it understand the world you’re building." Once the AI has a baseline understanding, developers specify the testing objectives.
These can range from traditional test cases—such as verifying menu navigation or collision detection—to more exploratory scenarios that are specially crafted to exploit the AI’s strengths. "Here you still need human expertise to shape the prompts and guide the AI toward the right behaviours," Holmgard adds. "The quality of the output depends heavily on how well you phrase the instructions." During a test run, the system records logs, captures short video snippets (referred to as "evidence"), and compiles textual descriptions of any anomalies it detects.
This package is then handed to a seasoned QA specialist who reviews the findings and decides the appropriate next steps. "We receive the bug description, the video proof, and the full log stream," Holmgard says.
"A human QA expert evaluates each report, confirms whether it is truly a defect or an intended feature, and provides feedback that the AI can learn from. Over time, this feedback loop refines the model’s ability to distinguish bugs from design choices." One of the most challenging aspects of automated testing is the subtle line between a bug and a feature.
The continuous feedback from human testers helps the AI internalize those nuances, gradually improving its judgment. Side envisions a future where QA personnel evolve into AI engineers, shifting the value chain upward. "Our people can move beyond repetitive test execution to become AI specialists," McBride says. "That not only raises their salaries but also opens new career pathways and potentially brings jobs back to regions that lost them due to low‑cost testing models." According to McBride, AI will free testers from mundane, repetitive tasks, allowing them to focus on strategic analysis and creative problem‑solving.
"Every new build triggers smoke tests and countless verification steps that bog down a tester’s workflow. By automating those routine checks, we let our staff engage in higher‑order thinking, which makes their careers more fulfilling." Benedikte Mikkelsen, Modl.ai’s chief product officer, echoes this sentiment: "There are many tedious test flows that a human can repeat endlessly without adding value. By offloading those to AI, testers can concentrate on evaluating gameplay feel, balance, and innovative testing approaches—activities that truly leverage human insight." McBride stresses that AI is not a replacement for human QA but a tool to elevate the workforce. "Testing rates in some low‑cost regions are so low that you can’t fully automate away the need for people.
Providing our teams with AI‑enhanced toolsets lets them extend their careers, move up the value chain, and stay with us longer." He also describes Side’s philosophy on AI as a "sliding scale"—a spectrum where human involvement never disappears. "That’s our ethos. The AI handles the repetitive, error‑prone parts, reducing mistakes and freeing humans to focus on the creative and analytical dimensions of QA." When asked about the types of games the solution currently supports, COO Oliver Deslandes notes that the technology is best suited for simpler titles—puzzle games, card games, or any experience with repeatable mechanics. "Fast‑paced 3D open‑world or fighting games are still a ways off.
We’re targeting clients who are willing to pilot this approach and understand that the technology will mature over time." Deslandes concludes, "It’s a journey. If you join us now and run a pilot, you’ll see how the AI evolves and where it can take your QA processes.
The potential for cost savings and quality improvements is significant, but it requires partnership and patience as the system learns."