Cubeia, a Swedish software company, is advancing in its journey towards AI-assisted development, reaching a pivotal third phase. The firm has transitioned from a laissez-faire use of AI by developers to establishing a structured AI-driven development process. This evolution prompts a significant inquiry: what should the organization focus on now that coding is no longer the primary limitation?
Stefan Grenstad, the COO of Cubeia, affirmed the company's commitment, stating, “We’re doing this 100%,” regarding their AI-driven approach to development. The emphasis has shifted to leveraging the additional capacity that AI has provided. Initially, six months prior, Cubeia was experimenting with substituting human-written code with AI-generated alternatives. However, the scope has since broadened to encompass a revised development pipeline, an evolved role for developers and quality assurance (QA) teams, and a transformed relationship with clients.
In its initial phase, Cubeia adopted a flexible approach to AI, allowing developers to utilize it at their discretion. Phase two introduced structure, standardizing the use of AI agents and creating a shared pipeline for all employees. This restructuring required Cubeia to address issues of quality, reliability, and agent collaboration, while also familiarizing staff with this novel method of operations. Grenstad remarked that the transition has largely been successful.
He highlighted a notable increase in productivity, stating, “During the hybrid period in Q1, Cubeia solved 259 issues. Once it moved to the AI-driven process, that figure rose to 421 – a 62% increase. Larger projects increased from 17 to 58. So the answer is yes: moving to AI-driven development has increased our output tremendously.”
The pressing question now revolves around how Cubeia will utilize this newfound 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,” Grenstad explained. This means teams must increasingly appreciate not just what they are building but why it matters and the value it generates.
Paul Crisp, head of marketing at Cubeia, emphasized the importance of measurement in this context. He noted, “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.”
An essential aspect of Cubeia’s new direction is being informed by customer input. Grenstad elaborated on this by stating, “We [recently] started asking, ‘What does this mean for the customer? What can the customer actually do with this?’” He recounted how customers expressed interest in employing Cubeia’s platform to enhance their own AI tools, such as building a custom casino landing page that integrates with Cubeia's APIs and customizing a player account management system for tailored bonuses.
This strategy focuses on enabling customers to access Cubeia’s platform rather than merely offering them additional AI tools. Presently, Cubeia is piloting the landing-page concept with its clientele. Grenstad remarked that these developments exceeded their initial expectations from six months prior, asserting the company’s readiness to accommodate customer-driven AI applications: “We’ve realized we’re going to be part of it.”
This surge in productivity has led Cubeia to reconsider its development team structure. The company is testing a rapid-response team dedicated to smaller customer needs alongside a separate team focused on larger projects and long-term strategies. While the current teams consist of five and eight members, Grenstad envisions potentially downsizing to two or three-person groups.
Team members can shift between rapid-response responsibilities and long-term projects based on their interests and areas where they can provide the most substantial contributions. However, increased capacity does not eliminate pitfalls. As AI agents generate multiple workflows concurrently, Cubeia is tasked with determining how much work a single individual can effectively manage and review.
“The bottleneck becomes the person reviewing everything,” Grenstad stated. He clarified that while code review remains vital—particularly for critical system components—there is a reduced need for exhaustive manual code scrutiny across the board.
This paradigm shift prompts Grenstad to rethink recruitment strategies. For roles centered on the player journey, candidates with domain-specific knowledge will be favored over traditional Java developers. “I would not necessarily prioritize senior Java or front-end specialists in the same way as before,” he expressed. With evolving needs, Grenstad underscored the significance of technical skills while highlighting the growing value of industry knowledge: “I’d rather take someone who’s good in the domain but doesn’t know any Java.”
This evolution raises questions regarding the pathway for future developers in the industry. With the reduced demand for traditional junior Java developers, the conventional ascent from junior programmer to senior engineer is becoming less apparent. Grenstad commented on the implications: “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.”
Additionally, Cubeia is reshaping its development framework by integrating QA earlier in the project lifecycle. In the traditional structure, roles were clear-cut: product owners defined what was built, developers devised how it was constructed, QA conducted tests, and operations managed the release. Now, QA specialists are involved from the onset, identifying automation opportunities and bridging coverage gaps as part of an ongoing process rather than culminating in a final stage. This transformation extends responsibility beyond problem identification to encompass the entire lifecycle of development, release, and outcome measurement.
Despite these changes, some staff members find adjusting to the new system challenging, especially those who derived satisfaction from traditional coding roles. Grenstad acknowledged this sentiment, stating, “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 important part of their work that they genuinely loved is now gone.”
Moreover, Cubeia faces strategic uncertainties regarding its reliance on AI providers. Currently dependent on Claude, Grenstad voiced concerns over potential shifts in pricing or quality from AI partners. Yet, he maintains an optimistic stance regarding Cubeia’s capacity to navigate these challenges, expressing, “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.”
Initially, Grenstad had a straightforward vision for Cubeia's AI initiative: “When we started this journey, my only vision was: ‘Let’s not be writing code in August.’” While Cubeia aimed to remove coding as a bottleneck in software development, it has learned that coding represents just one limitation in a much broader system. As the constraints shift, so too do the organizational demands, emphasizing the need for insights into domain knowledge, product understanding, quality assurance, prioritization, and customer needs. Consequently, the technology's role has diminished in significance relative to Cubeia’s strategic direction moving forward. The company’s focus has now turned to how it will wield this technology responsibly and effectively in the future.
