Side, a veteran services provider formerly known as PTW, has entered into a six‑month memorandum of understanding (MOU) with Denmark‑based artificial‑intelligence specialist Modl.ai. The agreement commits the two companies to co‑develop a quality‑assurance (QA) platform that leverages AI to streamline testing while keeping human expertise at the core. Modl.ai, a relatively new player in the gaming sector, was founded in 2018 and secured $1.7 million in seed capital two years later. Side, on the other hand, has a long‑standing reputation for delivering outsourced services across a range of industries.
By joining forces, the partners aim to combine Side’s deep QA know‑how with Modl.ai’s cutting‑edge machine‑learning capabilities to create what they describe as a “unique solution” for the games market. Martin McBride, Side’s chief commercial officer, explained that the collaboration is intended to deliver a "crown‑jewel" experience for developers: an AI‑enhanced testing workflow that can augment, rather than replace, traditional QA.
He noted that many attempts to embed AI in testing have fallen short because they required developers to hand over entire builds or even source code, creating friction and security concerns. Modl.ai’s earlier offerings relied on integration plugins and SDKs that demanded deep access to a game’s internals, a model that the company’s founder, Christoffer Holmgard, admits proved cumbersome over time.
The new system takes a very different approach. Instead of needing source code, the AI works with the same product a human tester would receive – the compiled game installed on a device that mirrors the end‑user environment. The AI runs the game, watches the screen, simulates inputs, and records every observable event.
A second AI layer then analyses this data, flags potential defects, and escalates them to a human reviewer for validation. The first phase of the workflow is what Modl.ai calls “building game knowledge.” Developers upload assets, design documents, screenshots, and other relevant materials, then converse with the AI in natural language to teach it the specifics of their title. Holmgard emphasizes that this interaction is deliberately conversational: "We just chat with the system about our game, adding images or design docs as needed," he says. This knowledge base allows the AI to understand the context of gameplay mechanics and narrative elements before testing begins.
Once the AI has a sufficient grasp of the game, developers define the test scope. This can range from classic regression scenarios to novel, AI‑friendly test cases that exploit the system’s strengths. Holmgard stresses that human input remains essential at this stage to craft effective prompts and guide the AI’s behavior.
"You need expertise to translate testing goals into language the AI can act on," he notes. During a test run, the AI captures textual logs, short video clips (referred to as “evidence”), and comprehensive telemetry.
These artefacts are compiled into a bug report that is handed to a seasoned QA specialist. The specialist reviews the findings, determines whether each issue is truly a defect or an intended feature, and provides feedback that the AI uses to refine its future judgments. Over time, this feedback loop helps the system distinguish subtle design nuances that might otherwise be misclassified.
Holmgard points out that one of the toughest challenges for any automated tester is discerning a bug from a deliberate game mechanic. By incorporating human feedback, the AI gradually becomes more attuned to the developers’ design philosophy, reducing false positives and improving overall accuracy. Side sees the partnership as a pathway to elevate its workforce. McBride explains that QA staff can transition from repetitive manual testing to roles akin to AI engineers, overseeing prompt engineering, data validation, and system tuning.
This shift not only raises salaries and career prospects but also opens the possibility of bringing jobs back to regions where low‑cost testing has been outsourced. "When we move people up the value chain, we can reinvest talent in markets that have historically lost these positions," he says.
Both companies argue that AI will make QA careers more fulfilling by automating monotonous tasks such as smoke tests and routine verifications. Testers will be freed to focus on strategic analysis, creative exploration of gameplay, and higher‑order problem solving. Benedikte Mikkelsen, Modl.ai’s chief product officer, adds that eliminating repetitive flows lets humans evaluate the fun factor and perform more inventive testing, which is where real value lies.
Importantly, McBride reiterates that AI is not a replacement for human testers. The price of low‑cost testing in certain regions remains a barrier to full automation, and human oversight is still required to interpret ambiguous results.
The company’s philosophy is that AI operates on a "sliding scale"—the more sophisticated the AI, the more it assists, but a human will always remain in the loop. "That’s our ethos," he says. "The tool lets us capture the best of AI while reducing error and fatigue, ultimately raising the quality of our output." When asked about the types of games the solution currently supports, COO Oliver Deslandes admits that the technology is best suited for simpler titles—puzzle games, card games, and other experiences with repeatable mechanics. Fast‑paced 3D open‑world or fighting games are still beyond the immediate scope and will require further development.
The team is targeting early adopters willing to pilot the system, acknowledging that results will vary based on project complexity and client expectations. In summary, the Side‑Modl.ai collaboration aims to create an AI‑enhanced QA pipeline that respects developers’ need for security, leverages natural‑language interaction to build game understanding, and keeps human experts in charge of final judgment. By doing so, the partnership hopes to reduce repetitive workload, improve defect detection, and open new career pathways for QA professionals, all while navigating the current technical limits of AI in gaming contexts.