Do CFOs Need to Learn to Code?

Writing code is yesterday’s problem

For a long time, “learning to code” meant memorising syntax, spending weeks on tutorials, and hunting for misplaced brackets and semicolons. That was tedious – and for most finance teams, far removed from the day-to-day.

That barrier has largely disappeared. Today’s AI tools handle the actual code writing. What remains is the more important skill: thinking one level higher – logically, structurally, analytically. Understanding how processes translate into software, what the building blocks are, how they fit together, and what can be automated.

It has never been easier to develop this mindset – because we can learn it together with AI.

Learning alongside AI

The best way to develop technical thinking is not a course – it is building something together with an AI tool. You describe what you need. The AI asks clarifying questions. You discuss alternatives. And step by step, you develop an understanding of how processes are modelled in software.

Tools like Lovable or Claude Code make exactly this possible. Lovable turns a plain-language description into a working web app – without a single line of code. Claude Code explains, asks follow-up questions, and suggests alternatives – and it is built right into the Claude app as the third tab, included in an existing subscription. Both pick up where weeks of onboarding used to be required.

Building apps – to understand the thinking

With Lovable or Claude Code, you can actually build a working prototype today – even without programming skills. A small dashboard, an automation script, an internal app. That is impressive – and educational.

But the real value is not the app itself. It is that, through building, you understand how software works: What is a data model? Why does a process need error handling? Where does complexity arise?

For production use – stable, secure, performant – you then need developers. But conversations with the IT team go very differently when you have traced the underlying logic yourself.

Conclusion: Not coding – but understanding what can be automated

Learning to code no longer means mastering a programming language. It means understanding which processes can be automated – and how to make that happen in collaboration with technology and AI.

That does not require a single line of Python. But it does require the willingness to engage with the logic behind software. And there has never been a better time to start than now.