Microsoft executives reveal how to use artificial intelligence to reshape the software development process Microsoft said that the role of artificial intelligence (AI) in software development is no longer limited to writing code, and the company is trying to fundamentally reshape the way engineering teams build software.

Microsoft executives reveal how to use artificial intelligence to reshape the software development process Microsoft said that the role of artificial intelligence (AI) in software development is no longer limited to writing code, and the company is trying to fundamentally reshape the way engineering teams build software. Charles Lamanna, executive vice president of Copilot, Agents and Platforms at Microsoft, detailed how Microsoft is deeply integrating AI into development workflows while maintaining a focus on customer success metrics. Lamanna emphasized that his team’s biggest milestone this year is not the launch of a flagship product or feature, but to completely change the way software is built through emerging AI technologies such as agent coding. He pointed out that this new working model will not be limited to coding or specific departments. Teams that fundamentally restructure their development processes around AI will perform far better than those that simply apply AI to existing processes. In this model, software engineers no longer need to hand-write code line by line, but can complete the development work by guiding the AI ​​agent system. Engineers' energy is freed to focus on more critical decision-making links such as strategic layout, architectural design, and quality control. Product development is no longer limited by coding ability; the bottleneck of the entire process has shifted to making the right decisions to create compliant products. Ramana said that although AI gives engineers more time to develop more features, he is guiding the team to keep the number of original feature development unchanged and instead use the saved time to focus on the reliability, performance and security of the software. This initiative has brought significant results, such as a double-digit decrease in the number of Power Automate-related service tickets year-on-year. He emphasized that customer success metrics and key performance indicators (KPIs) should be aligned whether AI is introduced or not. According to data from the relevant internal engineering team, after adopting agent coding, the team's failure events decreased by approximately 20%, support cases decreased by approximately 25%, and key upgrade matters decreased by approximately 30%. At the same time, the speed of accessibility work was also significantly improved. The executive concluded that engineering teams should not force AI into existing processes in order to use it, but instead need to reshape work habits around the technology. Teams that successfully transform should play a role model and encourage more departments to actively embrace AI. Based on this approach, Microsoft itself has successfully shortened the cycle of products from the concept stage to the mass market to a few months, which was unimaginable just a few years ago. However, LaManna also admitted that Microsoft has not solved all the problems yet. With the introduction of AI, how to operate in the huge