Cubeia, a Swedish software company, is making significant strides in its journey towards AI-assisted development with the introduction of a structured AI-driven process in its third phase. The question now looms: how should the organization evolve when coding is no longer the primary limitation?
COO Stefan Grenstad has affirmed, "We’re doing this 100%." The focus has shifted from simply substituting human-written code with AI-generated code to a more expansive view of development, encompassing a new pipeline, redefined roles for developers and quality assurance (QA), revamped team organization, and an evolving relationship with customers.
Initially, Cubeia's approach to AI was open, allowing developers to utilize it as they pleased. The subsequent phase introduced structure, aligning all developers with the same AI agents and processes. This shift prompted Cubeia to address challenges regarding quality, reliability, and inter-agent collaboration, while also helping employees adjust to the new workflow. Grenstad believes significant progress has been made in this area.
During the hybrid period in the first quarter, Cubeia managed to resolve 259 issues. Following the transition to the AI-driven process, this number surged to 421, marking an impressive 62% increase. Additionally, the scale of larger projects expanded from 17 to 58. As Grenstad states, “Moving to AI-driven development has increased our output tremendously.”
Now the pivotal question is how to utilize this increased capacity. Grenstad noted, “Because we’re not spending as much time coding, we’re spending more time on the business: talking about value and understanding the domain.” Teams are evolving from merely fulfilling requests to grasping the underlying reasons for developments and assessing their potential business value.
Paul Crisp, head of marketing at Cubeia, emphasizes the importance of measurement: “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.”
The direction of development is increasingly guided by customer input. Grenstad explained, “We [recently] started asking, ‘What does this mean for the customer? What can the customer actually do with this?’” This inquiry gained traction when customers approached Cubeia, requesting to utilize AI agents with their data streams on the Cubeia platform.
One customer, for instance, created its own casino landing page and sought to integrate it using Cubeia’s APIs. Another wished to enhance Cubeia’s player account management system for personalized features. The objective now focuses less on delivering another AI tool but on enabling customers to connect their existing AI infrastructures to Cubeia’s data and platform. A pilot project is already underway based on the landing-page integration.
Six months ago, such developments weren't on the immediate horizon, but Grenstad now views Cubeia as being positioned to “say yes” when customers present products developed with their own AI solutions. “We’ve realized we’re going to be part of it,” he acknowledges.
Organizationally, the boost in capacity has prompted Cubeia to rethink its development structure. The company is trialing a rapid-response team dedicated to smaller customer requests and incidents, alongside a more extensive team focused on larger projects and long-term strategy. Currently, these teams consist of five to eight members, though Grenstad envisions a future where teams could be streamlined to two or three people.
Developers will have the option to shift between rapid-response work and longer-term projects, depending on their interests and contributions. However, such increased capacity does not abolish bottlenecks entirely. With AI agents generating multiple streams of work simultaneously, Cubeia must navigate the balance of individual workloads. “The bottleneck becomes the person reviewing everything,” Grenstad explains, while also making it clear that sensitive system areas will continue to undergo careful scrutiny.
The transition to AI has also influenced Grenstad’s hiring practices. For example, a team focused on the player journey will seek candidates with knowledge of casinos and iGaming, moving away from a strict focus on traditional technical skills. “I’d rather take someone who’s good in the domain but doesn’t know any Java,” Grass stated.
This approach raises questions about the pathways for future developers. As Cubeia shifts away from dependency on standardized roles like traditional junior Java developers, the conventional career trajectory from junior to senior becomes less defined. “How do we avoid ending up with lots of old Java developers and nobody who understands Java because the juniors were never hired? That’s a super-interesting question. How do we fill up with younger people over time? That’s something we’ve been discussing.”
The company is also rethinking the traditional workflow where developers build and then pass their work to QA. Grenstad describes the prior model where product owners determined project scope, developers implemented it, and QA tested before operations finalized the release. Now, QA is integrated throughout the process, with product engineers automating aspects from the planning stage while QA specialists enhance automation and identify coverage gaps.
Responsibility for each project increasingly encompasses the entire journey from problem identification to construction, release, and outcome measurement. Although the transformation is promising, not all concerns have dissipated. Some coders lament the loss of opportunities to refine their skills through coding.
"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,” says Grenstad.
Cubeia also faces strategic challenges surrounding its reliance on AI providers. Currently dependent on Claude, Grenstad expresses concern over possible price increases or shifts in quality from AI vendors. Despite these uncertainties, he remains optimistic that Cubeia can learn to navigate these challenges as they arise. “We’re confident that the question marks will be solved as we work through them. We’ll see the problems, learn how to manage them, and adapt.”
In its initial stages, Cubeia embarked on its AI journey with a straightforward goal. Grenstad reflects, “When we started this journey, my only vision was: ‘Let’s not be writing code in August.’” The focus was on removing coding constraints, only to discover that coding was just one element in a much larger system. As bottlenecks shift, the organization’s demands will evolve towards a greater emphasis on domain knowledge, product development, prioritization, and understanding customer needs. Ultimately, the technology has become less central to the narrative, while the future use of technology has taken center stage.
