Cubeia Reports 62% Productivity Increase with AI-Assisted Development
Cubeia has boosted its development team's output by 62% following a full transition to AI-assisted coding. The company emphasizes that the shift required significant organizational changes rather than technical ones.
Software company Cubeia has reported a 62% increase in development issue resolution following its complete adoption of AI-assisted development processes. The transition, however, proved to be more about organizational restructuring than technological implementation.
According to COO Stefan Grenstad, the primary challenge was not in coding itself, but in adapting the surrounding processes and team structures. "When coding stops being the bottleneck, many assumptions about how a software team operates no longer hold true," Grenstad stated.
Cubeia implemented AI tools in three phases: unrestricted adoption, structured processes with unified agents, and finally, treating AI-assisted development as an organizational issue rather than a tooling one. This last phase involved committing fully and then determining the necessary business adjustments.
The company documented a rise in issue resolution from 259 to 421 within comparable periods. Additionally, the number of large projects in progress grew from 17 to 58. Cubeia cautions that increased throughput does not automatically equate to better overall output, noting that the bottleneck has shifted from coding speed to earlier stages in the development cycle.
Significant organizational changes include splitting teams into a five-person rapid-response unit and an eight-person long-term project team. Code review processes have also been adapted, with a greater emphasis placed on domain knowledge and directing AI effectively, rather than solely on manual coding proficiency.