The impact of AI on software development productivity

Jun, 02, 2025 | Read 3 min.

Since the launch of ChatGPT a couple of years ago revolutionized the world as we knew it, the implementation of Generative Artificial Intelligence in multiple industries has been growing exponentially. One of the sectors that has most embraced this change is, without a doubt, the technology sector. According to recent studies, 72% of software engineers already use it in their development processes, and of these, nearly half (48%) use it daily. This is having a major impact on how technical teams work and how their productivity is evolving. Or at least that’s what we might expect—but are we really becoming more productive with AI?

And what do we mean by productivity? That’s the first point we need to clarify before we can analyse whether AI is truly improving it. Real productivity means delivering business value faster, with fewer errors, and in a way that is sustainable for the teams.

So, the goal is to do more and better, without burning out your teams along the way.

Generative AI Comes to Help, Not Replace

Several tools have already established themselves in the market as “intelligent” assistants for software development. Among them, we find GitHub Copilot, Amazon CodeWhisperer, and Windsurf (formerly Codeium). It is also increasingly common to work directly with LLMs like Anthropic’s Claude 3.5 Sonnet or OpenAI’s GPT-4o, capable of generating, analysing, and improving code with remarkable accuracy.

These tools are helping to automate different tasks within development projects, such as:

Automatic code generation: They create snippets, functions, or even full modules from natural language descriptions—ideal for repetitive logic or boilerplate code.
Smart autocompletion: They can suggest lines or blocks of code in real time based on context, as GitHub Copilot does.
Automated testing: They generate unit tests and full test cases, improving test coverage and quality.
Assisted debugging: They analyse errors and propose fixes automatically, accelerating bug resolution.
Automatic documentation: They write technical descriptions, comments, and examples directly from the source code.
UI/UX prototyping: They turn textual descriptions into wireframes or visual interfaces in a matter of minutes.
Content management: They automate the creation of messages, documentation, UI text, and other materials.
Refactoring and optimization: They recommend structural and performance improvements to keep the code clean and efficient.

If teams are supported with this kind of assistance throughout all stages of the development lifecycle, it makes sense to think they’re now working more productively, allowing them to focus on higher-value areas like business logic, user experience, and other strategic decisions. But… is that really the case?

Does it really work?

Only data can tell us whether something is working, so let’s have a look at what some recent studies say about the impact of generative AI on software development productivity:

According to BairesDev’s study (2024) on the use of GenAI in a sample of more than 500 engineers:
- 23% of engineers using GenAI report a productivity increase of 50% or more.
- Another 71% say their productivity improved between 10% and 25%.
- Only 6% saw no change.
- The roles seeing the biggest productivity gains include DevOps engineers, Site Reliability Engineers (SREs), GIS specialists, and Scrum Masters, with increases of up to 50%.
- Work quality also improves: 74% of engineers say GenAI has increased quality, and over half report improvements between 10% and 25%.

A GitHub experiment with Copilot revealed that developers completed tasks 55% faster than those who didn’t use the tool.

And according to Google Cloud’s 2024 DORA report, teams with significantly higher adoption of AI tools, about 25% more than average, reported:
- A 2.1% increase in individual productivity
- A 3.4% improvement in code quality
- A 7.5% increase in documentation quality
- A 3.1% acceleration in the speed of code reviews

As we can see, the data is clear: AI not only speeds up work, but improves its quality and shortens development cycles.

With intelligence: the human factor still matters

If the data says what it says, organizations should already be verifying those improvements. This is where tools like Quanter can make a real difference, helping you measure and identify whether the improvements associated with AI adoption are actually being implemented and whether they’re delivering value to your projects.

Because the reality is different: although many organizations are already betting on generative AI, most have yet to implement it in a strategic and measurable way. According to a recent study by Hyperscience, the gap between excitement about generative AI and its effective application is significant. Why is that? What new risks do we face when implementing it?

We’ll explore those risks and how to avoid them in upcoming articles. But here’s a preview of one key insight: it’s not about using artificial intelligence “just because” but about integrating it where it truly solves a problem.

Will you join us on this journey?

Fuentes: