Cubeia, a Swedish software company, has embarked on an advanced phase of its journey towards AI-assisted development, shifting from a permissive use of AI by developers to a structured AI-driven development process. This evolution poses a critical question for the company: how will its operations adapt when coding is no longer the primary limiting factor?
Stefan Grenstad, the COO, confirmed a commitment to this direction, stating, "We’re doing this 100%." The new strategy emphasizes leveraging the extra capacity created by AI. Just six months prior, Cubeia focused on substituting human-written code with AI-generated alternatives. Now, the company is embracing a broader framework that encompasses a revamped development pipeline, new developer and quality assurance (QA) roles, reorganized teams, and an evolving relationship with customers.
In its initial phase, Cubeia took an open stance towards AI, permitting developers to integrate it as they wished. The second phase introduced a more structured approach, with all team members utilizing the same AI agents and following a standardized pipeline. This transition required Cubeia to address concerns regarding quality and reliability while ensuring that employees adapted to the changes. Grenstad believes that this groundwork has largely been laid.
"During the hybrid period in Q1, Cubeia solved 259 issues. After moving to the AI-driven process, that number climbed to 421 – a 62% increase. Large projects escalated from 17 to 58. So the answer is yes: our shift to AI-driven development has tremendously increased our output," Grenstad remarked.
The pressing query now is how Cubeia will utilize this enhanced capacity. "Because we’re not spending as much time coding, we’re spending more time on the business: discussing value and grasping the domain," Grenstad noted. Teams are transitioning from a mere execution of requests to a deeper understanding of the purposes behind their projects and the value they should generate.
Paul Crisp, the head of marketing, emphasized the need to measure outcomes effectively. For instance, he questioned, "If the objective is 10% more traffic, how do we measure it?" Establishing baselines, implementing changes, and evaluating results become crucial elements of their new approach.
Customer input has begun to steer Cubeia's development trajectory. Grenstad explained that the team is now exploring what customers want to achieve with AI. This shift became apparent when clients requested to connect their AI agents to Cubeia’s data streams and platform.
One customer created its own casino landing page, seeking to integrate Cubeia’s APIs. Another intended to use Cubeia’s player account management system for customized functionalities, such as personalized bonuses. Grenstad clarified that the goal is not simply to provide AI tools but to make Cubeia's platform compatible with the AI frameworks customers are creating. Currently, the company is running a pilot program based on the landing page initiative.
Reflecting on the future, Grenstad stated that Cubeia is preparing to affirmatively respond to customer inquiries about employing their own AI products with Cubeia’s system.
Organizational changes are accompanying this rise in capacity. Cubeia is setting up a rapid-response team for smaller customer requests alongside a separate group dedicated to larger projects and long-term plans. Although current teams consist of five and eight members, Grenstad envisions a future where they could be as small as two or three.
Developers will have the flexibility to transition between rapid-response tasks and larger projects based on their interests and optimum contributions. However, the increased capacity has not entirely resolved bottlenecks. With multiple AI agents generating work simultaneously, the challenge remains to establish a manageable workload for individual employees. Grenstad remarked, "The bottleneck becomes the person reviewing everything. That does not mean we never review the code itself; we still scrutinize critical aspects for architecture, performance, and reliability, but we don't apply manual code reviews universally."
Moreover, this AI integration is prompting Grenstad to rethink recruitment strategies. For teams focused on player journeys, the necessity for domain expertise in casinos and iGaming is growing more prominent. He articulated a shift in hiring priorities, stating, "I wouldn’t necessarily prioritize senior Java or front-end specialists in the same way as before. While technical skills remain important, domain knowledge is increasingly valuable. I’d prefer a candidate who excels in the domain, even if they lack Java expertise."
This evolution raises questions about the industry’s future workforce. As Cubeia moves away from the standard model that heavily relied on traditional junior Java developers, the pathway from junior programmer to senior engineer appears uncertain. Grenstad posed a concern: "How do we avoid ending up with lots of old Java developers and no young professionals, since juniors were never hired? That’s a super-interesting question."
Changing the approach to development, Cubeia is also departing from the conventional process where a developer completes a project only to hand it off to QA. In the previous structure, product owners determined project scope, developers decided implementation methods, QA conducted testing, and operations managed deployment. The new approach integrates QA earlier in the process, encouraging product engineers to consider what can be automated and identifying manual testing needs from the planning phase. This shift expands the responsibility to encompass the entire lifecycle of a project.
Despite these changes, Cubeia is not devoid of challenges. Some developers miss the coding aspect that they once cherished, struggling with the shift in their roles. Grenstad acknowledged this sentiment, noting, "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, an essential aspect of their work that they enjoyed is now gone."
Strategically, reliance on AI providers remains a critical consideration. Currently, Cubeia is heavily dependent on Claude, and Grenstad is mindful of potential challenges if AI providers increase costs or alter quality. Nonetheless, he believes that Cubeia can manage this uncertainty as they navigate forward.
Reflecting on the journey, Grenstad recalled, "When we began this journey, my only vision was: ‘Let’s not be writing code in August.’" The aim was to eliminate coding as the bottleneck in software development, but this endeavor has revealed that coding was just one of many interdependent constraints. As the focus of constraints shifts, so too do the organization's needs, now pivoting towards domain knowledge, product insights, quality, prioritization, and customer understanding. Consequently, the technology itself has taken on a diminishing role in the narrative.
