Shaun Abram
Technology and Leadership Blog
Report Summary: 2025 DORA report
I’ve been a huge fan of the annual DORA (DevOps Research and Assessment) report since reading the Accelerate book that explains the DORA metrics and how to improve them.
The 2025 DORA report (pdf attached here) just came out and has a particularly interesting section on AI Capabilities, The DORA AI Capabilities Model. I summarize some of the points here that I found most interesting here, but the original is well worth a read.
For technology leaders, it is no longer a question of if they should adopt AI, but how to realize its value. AI is an amplifier in software development, magnifying the strengths of high-performing organizations and the dysfunctions of struggling ones. Returns on AI come not just from the tools themselves, but from a strategic focus on the underlying
organizational system: the quality of the internal platform, the clarity of workflows, and the alignment of teams.
AI is ubiquitous in software development
AI adoption has become nearly universal with 90% of the DORA survey respondents use AI as part of their work and 80+% believing it has increased their productivity.
Conversational AI chatbots are the most common vehicle for interacting with AI, followed by AI embedded within the IDE, including predictive text, and tab-to-complete code suggestions.
Still, despite the ubiquity, 30% of respondents have little to no trust in the code generated by AI.
AI adoption requires more than just tools
This indicates that successful AI adoption requires more than just tools. AI can lead to a host problems, everything from silently deleted tests, to obviously broken functionality, and even deleted production data. But it can also massively improve outcomes. AI’s effects on performance depends on the system in which the work takes place. Teams need to be taught how to critically guide, evaluate, and validate AI-generated work, rather than simply encouraging usage.
Suppose the fastest you’ve ever traveled is walking at four miles per hour, and someone asks you to drive a car at 50 miles per hour. Without practice and training, you will undoubtedly wreck the car.
When using AI, your software development practices must evolve. We need fast feedback loops – faster than ever – to match AI-accelerated code generation. More than ever, we need to be able to develop, test, and deploy software independently. And we need a climate for learning, especially given the idiosyncratic nature of AI and its rapid rate of advance.
AI Capabilities Model
The DORA AI Capabilities Model identifies seven foundational practices that amplify the positive impact of AI on organizational performance. The effect of AI use on outcomes like throughput, code quality, and team and organizational performance was consistently amplified by these seven capabilities.
1. A clear and communicated AI stance
A clear and communicated AI stance means being clear on how developers are expected and permitted to use AI, and recommendations on which AI-assisted development
tools to use. Support developers’ experimentation.
A lack of clarity on the organization’s stance on AI use in software development is likely to manifest in the form of developers acting:
- too conservatively, using AI less because they are afraid of overstepping the organization’s parameters of acceptable use; and
- too permissively, using AI in ways that they should not.
Neither of these cases is optimal.
Instead, communicated expectations and acceptability of AI in software can help foster developers’ trust in AI, assuage concerns about data privacy, and scale adoption of AI-assisted development tools across the organization.
As long as the AI stance is clearly articulated and communicated, organizations can yield greater positive outcomes from AI adoption in software development processes.
2. Healthy data ecosystems
Healthy data ecosystems refers to the quality, availability and (lack of) siloing of an organization’s internal data systems.
If an organization invests in high-quality, accessible, unified data ecosystems, then that can yield higher benefits for performance than with AI adoption alone.
3. AI-accessible internal data
AI-accessible internal data is obvious closely related to and predicated on point 2, and refers to the degree to which AI tools are connected to internal data.
While AI tools trained on a general set of knowledge help developers, to be more effective and produce higher-quality code, AI can be even more impactful when given access to internal data sources and company-specific context.
Put differently, maximizing the benefits of AI may require a deeper investment than simply procuring AI licenses; if AI can’t access internal company data, how useful can it really be?
4. Strong version control practices
Strong version control practices are foundational to high-performing software development teams, providing a systematic way to manage changes to code over time. With generative AI, where the volume and velocity of code generation are dramatically increasing, the importance of these practices is amplified, to mitigate its risks.
When using Gen AI to assist with your coding, use version control, frequent commits, and rollbacks to revert changes. This provides a safety net allowing you to experiment and innovate with confidence, while being able to back to a stable state without fuss if something goes wrong.
AI can help you produce more code, faster so it makes sense that you need to be able to more rapidly undo those changes too.
5. Working in small batches
Breaking down work into small batches that can be quickly tested and evaluated has been a long time best practice:
- A smaller number of changes in each commit
- A small number of commits in each release
- And frequent releases
Small batches have a positive impact for AI-assisted teams too.
6. User-centric focus
A user-centric focus refers to the degree to which teams think about the experiences of end users; prioritizing that user experience and understanding its connection to business success. Engineers should be treating user experience as a top priority, and creating value for users is key to the success of the business.
AI’s positive influence on team performance is amplified when teams adopt a user-centric focus. Without a user-centric focus, AI adoption is unlikely to help teams.
Organizations that encourage AI adoption will benefit from incorporating a rich understanding of users, their goals, and their feedback.
AI can make use faster. But are we just getting faster at creating features that don’t deliver value to our users?
7. Quality internal platforms
“Platforms” refer to a set of capabilities that is shared across multiple applications or services.
AI’s impact depends on organizations having quality internal platforms.
Organizations with quality internal platforms have a great positive influence from AI’s adoption.
Conclusion
Successfully leveraging AI in software development is not as simple as just adopting new tools. Rather, organizations must cultivate a specific technical and cultural environment to reap the greatest rewards.
Tags: AI, artificialintelligence, dora, dorareport, genai, summary
Subscribe to RSS Feed