Nathen Harvey

We have entered the era of AI. Developers around the world, and in your organization, are now using AI for everything. Of course, it's not just for developers. AI is being used by everyone across the software development life cycle.

Product managers are using AI to help write and prioritize user stories. Developers are using AI to help generate code, configuration, and tests. Operators are using AI to help troubleshoot and respond to customer issues.

The world is upside down. Everything has changed. We must throw out everything we've learned and embrace the new AI-powered world.

Whoa. Wait. I'm sorry. That's not right.

Look, things have certainly changed, and this AI moment is likely to have a lasting impact on the work we do every day. But we cannot simply throw out the lessons of the past.

For a decade, DORA has been researching the capabilities of high-performing, technology-driven teams. DORA provides a framework for measuring and improving software delivery performance. By utilizing DORA's Quick Check, teams can establish a performance baseline and pinpoint areas for improvement.

With this baseline in hand, a team can now begin to explore the capabilities that they'll need to improve. Broadly speaking, these capabilities support a climate for learning, fast feedback, and shipping small changes.

Artificial intelligence will play a crucial role in accelerating the software development lifecycle. In fact, in 2023, DORA's research found that teams with faster code reviews achieve 50% better software delivery performance.

Can AI contribute? I think so. I think AI can significantly help enhance code reviews in a number of different ways. For example: facilitating the creation of higher-quality code that's simpler to review; segmenting large change sets into smaller, more manageable chunks for review and deployment; explaining the code that we're reviewing; and verifying that code meets team standards, policies, and compliance controls.

Moreover, AI can assist in other crucial software delivery processes, such as enhancing code comprehension and reducing technical debt, automating the integration and testing of code changes, accelerating testing processes, and improving test data management.

While AI can help improve existing tools and processes, there are also things we can carry forward when it comes to building and managing applications driven by AI: smaller changes, fast feedback, and a climate for learning. These are all going to be important as we collectively discover effective ways to build, deploy, operate, and change AI models and the way that we interact with them.

Adopting an experimental mindset is essential. By setting performance baselines, running experiments with AI, and monitoring the impact on overall performance, teams can strategically implement AI solutions. When successful, continued investment and exploration of additional AI opportunities can further optimize the software delivery process.

By leveraging DORA metrics and AI capabilities, teams can pinpoint areas of friction, inject AI solutions, and achieve continuous improvement in their software delivery performance.

Learn more at dora.dev, and let's keep this conversation going right here at the Enterprise Tech Leadership Summit and beyond in the DORA community. I'll see you there.