Follow ZDNET: Add us as a favorite source On Google.
Key takeaways of zdnet
- Almost all developers now rely on the AI Tool.
- AI increases strength and increases dysfunction.
- High quality platforms AI are very important for success.
Google released it 2025 Dora Software Development ReportDora (Devops Research & Assession) is a research program of Google (part of Google Cloud Organization). Dora examines capabilities and factors that run software distribution and operational performance.
This year, the Dora Project surveyed 5,000 software development professionals in industries and chased with more than 100 hours of interview. This can be one of the most comprehensive studies of AI’s changing role in software development, especially at the enterprise level.
Also: 10 Chatgpt Codex Secrets I learned only after programming 60 hours with it
The results of this year are particularly relevant as AI has infiltrated software development to an extreme degree. Some encouraging notes are shown in the report, but also show some areas of the actual challenge.
In writing this article, I have gone through the 142-Page report and have drawn five major comments that cut through publicity to find out what is really changing in software development.
1. AI is now widely used in development
According to the survey respondents, somewhere between 90 and 95% rely on software development for work. The report mentions 95% in Intro and later 90% in an expansion section, but no matter what number you choose, almost all coders are now using AI. According to the report, it is a leap of 14% from last year.
The average time spent in interaction with AI was two hours per day. However, there are some more nuances for this. For example, only 7% of the respondents reported “always” using AI, when faced a problem to solve. The largest group, 39%, moves to “sometimes” AI for help. But what I was killed by using a full 60% AI “almost half time” or more when trying to solve a problem or complete a task.
Eighty percent of the programmer reported overall increase in productivity, but only 59% reported that their code quality has improved. Another major metric is this: 70% of the respondents rely on the quality of AI, while not 30%.
Also: I got 4 years product development in 4 days for $ 200, and I am still shocked
Share me a personal idea on it. I have just finished a huge coding sprint made possible by AI. The code that came out was almost never right on the first run. I had to spend a lot of time to do this to AI to get it right. Even once the work was done, I went back to a full QA sweep, where I found more errors.
My conclusion is that there is no way that I could get anywhere near the quantity of work that I did just without AI. But there is no way in Hek, I am going to rely on any code that AI writes too much review, verification and testing. Of course, it is not very different that when I was a manager and handed over to employees or contractors to coding.
2. Think of AI as an amplifier
This was one of the more attractive consequences coming out of the study. The Dora team says AI has become an amplifier. Essentially, AI “enhances the strength of high -performing organizations and the dysfunction of the struggling people.”
It makes so much understanding. If you read my most recent article on “10 Chatgpt Codex Secrets, I learned only after a 60 -hour pair of programming with it,” I told that AIS makes big mistakes quickly. A deformed signal may send an AI to destroy some major destruction. I had the experience where Codex decided to remove a large portion of one of my files, and then immediately checked in those changes in Github.
Also: I coded 24 days in 12 hours with $ 20 AI tool – but a big loss is
Fortunately, I was able to roll back those changes, but I noticed the mass work disappeared rapidly, as much as I could take a sip of coffee.
Essentially, the more effective and conducted a team, the more AI will help. The more a team is scattered or random, the more AI will get hurt. In my case, I have a truly good modification control practice, so when AI ate my homework, I was able to get it back because I had put my codebase in those controls before giving AI my first reach.
3. Seven team archtypes in AI era
So who wins and who loses? The Dora team identified eight factors, determining the overall performance of a team.
- team performance: A team’s effectiveness and collaborative power
- Product Performance: Quality and success of products are being produced
- Software delivery throughput: Speed and efficiency of distribution process
- Software delivery instability: Quality and reliability of distribution process
- Personal Effectiveness: A sense of effectiveness and achievement for individual team members
- Valuable work: Which individual team members feel that their work is valuable
- Conflict: How much do you get in the way of people who try to complete their work
- burn out: Emotions of tiredness and condolences among team members
He then measured these factors against the respondents and his teams. This helped identify seven team arctapes.
- Founder challenges: Survival mode, interval everywhere
- Inheritance: Continuous fire fighting, unstable system
- Binding by procedure: Stable but bureaucracy surprised
- High impact, low rhythm: Strong output, unstable distribution
- Stable and method: Intentionally speed, consistent quality
- Practical artist: Reliable, fast, moderately engaged
- Harmonious high acquisition: Permanent, stable, top performance
AI, the report says, there is a mirror of organizations. The use of AI makes the strength and weaknesses of the teams more clear. But what I found especially interesting is that “speed versus stability” is a myth.
The idea is that you may intensify or you can produce good codes, but not both. As it has been revealed, the top 30% of the respondents fall into the harmonious high-obtained or practical artists, and they are quickly producing the output, and the quality of that output is high.
4. Seven major practices
The report emphasizes, “Successful AI adoption is a system problem, not a tool problem.” Dora feels like number seven. They say that the following seven major practices run the effects of AI (for good or bad).
- AI policy: The clear, AI stance of an organization communicated.
- Data ecosystem: The overall quality of internal data of an organization.
- Accessible data: AI tools associated with internal data sources.
- version control: The systematic way to manage changes in the code.
- Small batch: Breaking changes in small, manageable units.
- User focus: Teams that prefer the experience of the end users.
- Quality platform: Shared capabilities are available in the organization.
As you can imagine, successful teams employ more of these practices. While unsuccessful teams can have highly productive individual programmers, it is a lack of these basic things that bring them down.
They recommend, “Consider your AI as an organizational change. The biggest return will come from investing in basic systems enhancing the benefits of AI: your inner platform, your data, your data ecosystem, and the main engineering discipline of your teams. These elements are the conditions required to measure AI’s ability.”
5. Two factors that affect AI success
Last year, it was quite a big news when the previous Dora report revealed that AI had actually reduced software development productivity, rather than enhanced it. This year, the opposite is true. The Dora explorers were able to identify two major factors that changed the results.
Development organizations are more familiar than AI and know how to work more effectively than a year ago. The study shows that 90% of developer organizations have adopted platform engineering. It is the practice of building strong internal development platforms that collect equipment, automation and shared services for a growth team.
Also: Best AI for coding in 2025 (and what not to use)
According to Dora, when the internal platform works well, developers spend less time to fight the system and make more time. If you see AI as an amplifier, you can see how good systems can actually improve results. Interestingly, if the platforms are weak, AI does not improve organizational productivity. Good internal platforms are a very clear condition for effective AI use.
The next factor sounds like a discussion out of a workplace SICST, but is really quite important. It is VSM (or Value Stream Management). The idea is that the managers make a map of how it works from work to delivery. It is basically a florchor for operation instead of bits only.
Looking at every step, teams can identify the problem areas, such as very long code reviews or stall release in different stages. The report states that the positive effect of AI adoption is “dramatically amplified” in organizations with a strong VSM practice. For records, the word “dramatically” appears in the report four times.
The report stated, “VSM AI acts as a force multiplier for investments. By providing a system-level approach, it ensures that AI is applied to the right problems, local productivity gains are transformed into important organizational benefits rather than only to make more downstream chaos.”
What does it all mean for software development
There are some clear conclusions from the report. First, AI has moved the mainstream from Hype to the mainstream in the Enterprise Software Development World. Second, the real benefit is not about the equipment (or even the AI you use). It is about the creation of solid organizational systems. Without those systems, AI has little benefit. And third, AI is a mirror. This indicates and enhances how good (or bad) you already work.
What do you think? Is your organization using AI Tool in software development? Do you see AI as promoting a real productivity or is there something that adds more instability? Which one of the seven team’s arcatypes is closest to your experience? And do you think that practices like platform engineering or VSM really make a difference? Share your thoughts in the comments below.
You can follow my day-to-day project updates on social media. Be sure to subscribe to My weekly update newsletterAnd follow me on Twitter/X @DavidgewirtzOn Facebook Facebook.com/davidgewirtzOn Instagram Instagram.com/davidgewirtzOn blue @Davidgewirtz.comAnd on youtube Youtube.com/davidgewirtztv,
