Cubeia’s road to AI-assisted code (part three): It’s no longer about the code

Swedish software company Cubeia’s journey towards AI-assisted development has entered its third phase as the company has moved from allowing developers to adopt AI as they see fit, to building a structured AI-driven development process.

It now faces a broader question: how should the organisation work when coding is no longer the main constraint?

“We’re doing this 100%,” COO Stefan Grenstad says of Cubeia’s AI-driven development process.

Now, the focus is on what the company does with the additional capacity AI has created. Six months ago, Cubeia’s experiment was essentially about replacing human-written code with AI-generated code.

Since then, it has evolved into something broader: a different development pipeline, a different role for developers and quality assurance (QA), a different way of organising teams and, increasingly, a different relationship with customers.

Creating business value

Cubeia’s first phase was an open approach to AI. Developers could use it whenever they wanted. Phase two brought structure, with everyone using the same agents and working through the same AI-driven pipeline. That required Cubeia to solve questions around quality, reliability and how agents could work together, while getting employees comfortable with the new way of working. Grenstad believes that work has largely been completed.

“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,” Grenstad says.

But the more important question now is what Cubeia does with that 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 says.

Rather than simply building what is requested, teams are increasingly expected to understand why something is being built and what value it should create.

As Paul Crisp, Cubeia’s head of marketing, points out: “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.”

‘Say yes‘

The next development is increasingly being driven by Cubeia’s customers. Now the company is asking what customers will want to do with AI, Grenstad explains.

“We [recently] started asking, ‘What does this mean for the customer? What can the customer actually do with this?’ And I think we would have continued looking for that answer if the customer hadn’t come to us and said: ‘Can we let our AI agent deal with your data streams and use your platform?'”

One customer built its own casino landing page and wanted to use Cubeia’s APIs to make it work with the platform. “Another wanted to use Cubeia’s player account management system to build its own functionality, including personalised bonuses,” Grenstad recalls.

The idea is less about Cubeia providing another AI tool than making its platform and data accessible to AI infrastructure customers are already building themselves. Cubeia is now running a pilot based on the landing-page use case.

This was not necessarily on the horizon just six months ago, Grenstad says, but now he believes Cubeia should ultimately be able to “say yes” when customers arrive with products built using their own AI tools.

“We’ve realised we’re going to be part of it,” Grenstad says.

A different development structure

The increase in capacity is also changing how Cubeia organises its development team. The company is experimenting with a rapid-response team focused on smaller customer requests and incidents, alongside another team working on larger projects and the longer-term roadmap.

The current teams have five and eight people, but Grenstad believes they could eventually become much smaller.

“In my dream, it would be two- or three-person teams,” he says.

Developers can also move between rapid-response and longer-term projects depending on their interests and where they can contribute most effectively.

But more capacity does not mean Cubeia has eliminated bottlenecks. With AI agents producing several streams of work in parallel, the company is still trying to find the right balance of how much work an individual can manage and review.

“The bottleneck becomes the person reviewing everything. That does not mean we never review the code itself. For sensitive or critical parts of the system, we still look more closely at the implementation, particularly from an architecture, performance, security and reliability perspective. But we no longer apply that level of manual code review to everything.” Grenstad says.

The shift to AI is also changing how Grenstad thinks about recruitment. A team working on the player journey, for example, needs people who understand casinos, iGaming and what a strong player experience should look like.

“I would not necessarily prioritise senior Java or front-end specialists in the same way as before,” Grenstad explains. “Technical skills remain important, particularly for people who can assess architecture and ensure systems remain reliable, but domain knowledge is becoming increasingly valuable. I’d rather take someone who’s good in the domain but doesn’t know any Java.”

That raises a question about how future developers will enter the industry. If Cubeia no longer needs front-end traditional junior Java developers in the same way, the traditional path from junior programmer to senior engineer becomes less clear.

“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.”

Big changes and unknowns

Cubeia is also moving away from the traditional process in which a developer builds something and then hands it to QA.

In the old model, Grenstad says, the product owner decided what to build, the developer decided how to build it, QA tested it and operations eventually handled the release.

As discussed last month, QA has become part of the process rather than a final stage. Product engineers consider from the planning stage what can be automated and what needs manual testing, while QA specialists now strengthen automation and identify gaps in coverage.

The responsibility for a piece of work increasingly extends from identifying the problem to building, releasing and measuring the result.

The transformation has not eliminated all of Cubeia’s concerns. And inevitably some coders are missing their original role.

“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.”

There are also strategic questions around dependence on AI providers.

Cubeia is currently heavily reliant on Claude, and Grenstad questions what could happen if AI providers raise prices or if quality changes. For now, however, Grenstad sees the uncertainty as something Cubeia can learn to manage.

“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.”

Sharing the journey

When Cubeia began their AI-journey, the objective was simple.

“When we started this journey, my only vision was: ‘Let’s not be writing code in August,’” Grenstad says.

Cubeia set out to remove coding as the constraint on software development, but in doing so, the company has discovered that coding was only one constraint in a much larger system. As the bottleneck moves, so do the demands on the organisation: towards domain knowledge, product thinking, quality, prioritisation and understanding what customers need.

The technology itself has therefore become less important to the story.

What Cubeia does with the technology moving forward has become the focal point.