3 ways AI is changing how we build software in 2025

What AI agents mean for the future of CI/CD and developer tooling.

AI was a significant catalyst for change last year. In 2025, organizations that have already started integrating AI into their processes will begin to see a return on investment, allowing them to refine and refocus their use of AI in software development.

Recent research by GitLab found that 78% of organizations either actively use AI in their development processes or plan to implement it in the next two years. The use cases are likely varied, reactive, application of AI for experimental and non-critical projects. This year, the most strategically-minded organizations will advance their AI strategy by driving engineering efficiencies and adding AI agents throughout their entire software supply chain.

3 trends shaping the software development in 2025

With the shift toward higher productivity in development, organizations will focus more on ROI and quantifying AI’s impact. Here are three trends shaping the future of software development in 2025.

Platform engineers will benefit from AI-driven efficiencies

As pattern recognition improves, AI can reduce friction from automating software releases to production. Teams can use AI to further platform engineering goals: Codifying training, policies, and checks and balances, helping to identify areas for improvement and reveal best practice.

Embedding AI into platform engineering will increase application development, letting organizations accomplish their goals more efficiently. By creating reusable building blocks that encapsulate common functions in software delivery, platform engineers can let non-technical team members assemble delivery pipelines with intuitive low-code techniques for testing, environment management, and release orchestration.

AI agents will be key to changing software supply chains

AI agents can change the software supply chain by automating and optimizing processes, from continuous integration to continuous deployment. The transition from fragmented AI applications into fully integrated AI-powered workflows will initially gain traction in open source ecosystems, like software libraries, where AI agents will likely be built and shared with the community.

As developers and organizations see the benefits of AI-driven automation in open source projects, we can expect a rapid expansion into commercial enterprise solutions. Internal development teams and platform engineers will increasingly be tasked with building, extending, and integrating AI agents in the software supply chain.

Data governance and cloud cost optimization will dominate 2025

Organizations will intensify the scrutiny of operational efficiencies and cloud spending. Companies will focus on ROI and total cost of ownership–conducting granular cost assessments at the application level–rather than prioritizing development speed. We’ll likely also see a rise in the practice of FinOps, which serves as a link between finance, product, and engineering teams. FinOps provides a framework for evaluating issues, identifying new opportunities for efficiency, and building remediation plans.

Companies will use this new foundation to compare an application’s revenue to development and maintenance costs, potentially accelerating the shift to on-premise or hybrid environments. The complex and costly nature of cloud-native modernization and increasing data privacy and AI regulations will emphasize the need for data control and governance. When technical operations align with financial objectives, organizations can ensure they receive the highest ROI for their cloud infrastructure and software development investments.

Conclusion

This year, the most competitive organizations will embrace AI strategically and intentionally. By incorporating frameworks to optimize cloud spending, using AI to drive efficiencies in the software supply chain, and using AI to level-up platform engineering efforts, we will see transformations and opportunities for innovation in organizations and across the broader technology ecosystem.

More from We Love Open Source

• What is Artificial Intelligence (AI) and the three things it does well
• AI isn’t scary—lack of knowledge is. Here’s how to stay ahead
• Discover Goose: Automate your developer setup with this AI agent
• Getting started with Streamlit

This article is adapted from “AI in software development: Looking beyond code generation” by Lee Faus, and is republished with permission from the author.