Cubeia, a Swedish software company, is advancing in its transition to AI-assisted development, now entering a third phase that emphasizes a structured AI-driven development process. This evolution raises a crucial question about how Cubeia will function when coding no longer presents the primary constraint in their operations.
"We’re doing this 100%," said COO Stefan Grenstad, affirming the company’s commitment to an AI-driven development approach. The shift is not merely about producing code but entails a deeper engagement with the additional capacity that AI enables. Six months prior, the focus was on substituting human-written code with AI-generated alternatives.
Now, the scope has widened to encompass a revamped development pipeline, altered roles for developers and quality assurance (QA) professionals, reorganized team structures, and a notably different rapport with customers. This transformation aims to enhance business value.
Cubeia's first phase embraced an open and flexible AI integration, allowing developers to employ AI tools as they wished. In the second phase, a structured workflow was adopted, standardizing the use of AI agents and a systematic pipeline. This transition involved addressing key concerns regarding quality and collaboration among AI agents while ensuring employees adapted to the new methodologies, a task Grenstad indicates is largely complete.
In the first quarter during the hybrid phase, Cubeia addressed 259 issues. Hi-technology processes skyrocketed this performance, with resolution numbers jumping to 421, marking a 62% increase, and larger project counts rising from 17 to 58. Grenstad stated, "So the answer is yes: moving to AI-driven development has increased our output tremendously."
The pivotal question now turns to how Cubeia will capitalize on this enhanced capacity. "Because we’re not spending as much time coding, we’re spending more time on the business: talking about value and understanding the domain," he explained.
This shift means development teams are beginning to inquire not just about what is to be built, but why it matters and what kind of value it should create. Paul Crisp, Cubeia’s head of marketing, emphasized the importance of measurement in determining success, noting, "If the objective is 10% more traffic, for example, how do we measure it? We need to establish a baseline, implement the change, measure it and see whether it goes up by 10%. If it only goes up by 1%, maybe we need another iteration because we haven’t fulfilled the objective."
Customer demand is increasingly guiding Cubeia's development trajectory. Grenstad remarked that they have started asking customers how they envision utilizing AI. One client developed a casino landing page and sought to integrate Cubeia’s APIs, while another aimed to create custom functionalities, including tailored bonuses, using Cubeia’s player account management system. The company is piloting its response to such customer needs, aiming to let clients use their own AI infrastructures along with Cubeia’s platform and data.
This shift was not initially anticipated but is now seen by Grenstad as an essential evolution, as he expressed a desire for Cubeia to be able to "say yes" when customers present AI-driven products built on their platform.
The increasing capacity is also prompting Cubeia to rethink its development team structure. They are exploring a rapid-response team to handle smaller customer needs alongside a core team focused on larger projects. Current teams comprise five to eight members, but Grenstad envisions the potential for smaller collaborative units of two to three people. This flexibility allows developers to shift between urgent requests and long-term initiatives based on their strengths and interests.
However, the introduction of AI does not entirely eliminate bottlenecks. With AI agents producing multiple outputs simultaneously, Cubeia continues to grapple with how much work one individual can realistically oversee. "The bottleneck becomes the person reviewing everything," Grenstad noted, insisting that scrutiny remains crucial for sensitive system areas, especially when it comes to architecture and security.
AI's integration also influences recruitment strategies. Unlike before, Grenstad said, there is now greater emphasis on finding talent that understands the iGaming domain rather than solely on traditional technical skills. "Technical skills remain important, but domain knowledge is becoming increasingly valuable. I’d rather take someone who’s good in the domain but doesn’t know any Java," he explained.
This raises broader implications for how future developers will be trained and introduced into the industry. As the need for entry-level Java developers diminishes, bringing in younger talent while ensuring knowledge continuity in Java presents an intricate challenge.
Cubeia is fundamentally rethinking its development model, steering away from the conventional handoff process wherein coding precedes quality assurance. Grenstad explained that the new approach integrates QA as an ongoing part of development, with product engineers considering automation from the planning phase onwards while QA specialists enhance automation and identify areas requiring more attention.
Yet, despite these advancements, some long-time developers miss the direct coding aspect they once enjoyed. "We still have people who loved the coding part – solving problems with code and writing beautiful code. They’re struggling with the change. For some people an important part of their work that they genuinely loved is now gone."
Concerns over dependence on AI service providers also loom large at Cubeia. Currently reliant on Claude, Grenstad voiced his apprehensions regarding potential price hikes or fluctuations in service quality. Nevertheless, he remains optimistic, asserting that as issues emerge, the company will learn to navigate them and adjust accordingly.
When Cubeia embarked on this AI journey, their singular aim was clear: "Let’s not be writing code in August," Grenstad recalled. They sought to eliminate coding constraints in software development but uncovered broader systemic limitations. As the bottleneck shifts, the corresponding demands on Cubeia evolve toward deeper domain expertise, product understanding, and customer engagement, illustrating that technology's importance is increasingly secondary to how the technology is utilized going forward.
