Training Developers for AI: One CIO’s Approach

Even as the hype over AI has washed over the IT industry, I have seen developers shy away from AI development aids for fear those tools would replace them. Neither the hype nor the fear is justified. AI development tools are not perfect and will not replace developers who stay on top of their game. But AI can save money and time today if used for the right purposes. Here’s how we have taken a practical, informal approach to training our developers in AI to help our business and their careers.

Should IT Workers Worry?

I’ve had developers in my organization who were genuinely worried that AI will make them redundant. Some of them were even afraid to use AI assistants. They would find reasons to avoid them, trying to prove they could do better work than the AI tools. And on an emotional level, I get it.

But this approach is futile because AI tools can help you be more productive. One example is using AI tools to write tests for code, which is time-consuming and routine work that many developers tend to skip.

Another example is bringing developers quickly up to speed in a new area such as a database. Instead of spending hours reading extensive documentation, they can ask an AI assistant for a specific solution to a particular problem. This way, they get the information they need more quickly. As with writing tests, this eliminates drudge work, saving time for more complex tasks.

This practical approach to AI actually improves their job security because if they don’t learn how to use AI tools today, tomorrow they may lose their jobs to others who have.

That does not mean AI is perfect. AIs, especially large language models (LLMs), are simply machines. They cannot distinguish right from wrong but merely reference the data fed to them (if they don’t forget it along the way) and perform the tasks set by their users. While AI can assist with specific tasks like optimizing algorithms or writing small, standalone pieces of code, and writing texts, it currently struggles when the context is too broad or is not structured and clearly defined.

Therefore, developers and IT professionals can sleep well, as long as they stay aware of how AI can and cannot help them and learn how to use it appropriately.

Offering AI Training the Practical Way

To get the most of AI development tools, this August we started offering hands-on AI training for the roles we feel can provide the most immediate value. These include software development, quality assurance, DevOps, and system engineering. But this training is also open to any software development staff, such as security engineers.

We are organizing it in the form of localized workshops where employees actively engage with AI tools. These sessions aren’t just about theory or abstract concepts but about practical application. These sessions are led by our own people who use AI coding assistance, including a developer, a technical lead and an engineering manager. During the workshops, participants generate code, debug issues, and optimize projects using AI in real-time. The goal is to demonstrate where AI can be effectively used and where it might not be as useful.

One area on which we focused was prompt engineering – how to phrase a question to the I tool to get the most accurate and useful response. What’s been interesting is watching the transformation. Developers who were hesitant at first, thinking AI might replace them, are now seeing how it can make them more efficient. They are not losing their jobs but becoming better at them.

The Benefits of Micro-Learning

I no longer consider formal IT training necessary. AI courses – as with most IT topics – often become outdated by the time they end, especially given how rapidly new AI iterations are introduced. Instead of relying on traditional training programs, we see greater value in practical micro-learning activities that address real-world problems rather than theoretical scenarios.

This approach to AI training, although informal, ends up saving company resources, not to mention time. Sure, the employees don't get a two-day session filled with an instructor covering general industry examples or a diploma with a beautiful stamp on it (to be tucked away in a drawer somewhere). What they do get is targeted training with practical, readily applicable examples and scenarios that ends up saving them (and the company) thousands of hours each month combined.

There is also the positive effect of being instructed by one of their own. Our developers don’t shy away from asking questions and can talk about what they have learned informally with their peers.

This training “by the devs, for the devs” ensures that the learning is relevant and immediately applicable, making it more effective than conventional, lengthy training programs. We use the same approach for sessions on general problem solving, such as adapting new technologies into existing projects.

We try to organize monthly internal online courses across our global locations that are recorded so there is no limit to how many people can participate. But even on an informal, daily basis we are seeing people sharing best practices. We encourage people to take the initiative and run their own small workshops within their teams. 

No More AI FOMO

I take a practical approach to AI. There is still a lot of hype around it, and companies have been caught up in a whirlwind of excitement and fear of missing out (FOMO). Both stem from uncertainty surrounding AI. It is still a new field, so fear is understandable.

But this is where the CIOs and CTOs need to step in – to encourage their teams to try out new things while making sure they remain within a practical framework. It is not enough to just adopt AI. You need to make sure it is adding value to your business and that your employees actually feel confident using it.

Ditch the fear. Invest in purposeful training and focus on showing people how AI can help both – their jobs and themselves. Only then will it bring results.