Work In Progress

Experiences with Claude Code

Coding for enjoyment

If you’re read any of the other articles on this site, you will have gathered that I enjoy coding. Whilst it is entirely true that the code itself may not be the most elegant that you’ve seen, it is also the case that writing it has been of great enjoyment. Even if, on the odd occasion, it has also been rather frustrating.

And then Claude Code came along.

Enter Claude Code

At work we’ve just released a new CLI, neo4j-cli ( Go look at neo4j.sh to learn more. It’s worth it ) This is a CLI that was not only built to work for people and AI, it was also ( mostly ) written by AI.

Four of us formed a team , invited Claude Code and shipped it in 8 working days.

Wearing of two hats

I’m a Product Manager by trade and that’s what I brought to the team. But , as I previously mentioned, I can write code but I’m much better at reading it.

What Claude Code did for me was to play to those two strengths. I could use my PM abilities to describe a problem with desired outcomes to Claude - the What - and then be able to read the code that was written - the How.

The code areas I touched were well within my comfort zone. And relatively minor. I was able to effectively use Claude Code to describe what was needed and then check the produced code. Only the slightest of oversight from my much more talented peers was required. Anything more would have slowed us down.

Product Manager Engineers?

I have read this title before. There’s chatter on this topic , AI enabling Engineers to be more like Product Managers and vice-a-versa.

I have always thought Product Managers having an understanding of Engineering is part of knowledge base toolkit that PMs have. It sits comfortably alongside other tooling such as what our customers need, what our business needs, market conditions and many others. AI has the pontential to morph that Engineering PM tool and how it could be used.

I say could rather than should as we need to consider the latter carefully. For me Claude Code and other AI tools allows discovery to go deeper and further than before and not just for classics such as locating customer pain points. I can now undertake prototyping with the goal of having better defined Asks for Engineering. As an example I recently contributed a Go SDK for the Aura API ( You can find that at https://github.com/neo4j-contrib/aura-go-sdk ) with the intention of exploring how Developers can use SDKs for Aura API and if they actually want to!

Would I go beyond this? I think not. Coding is a hobby for me. Tools like Claude Code up-level my coding ability to become a useful PM discovery tool .

And that - for me anyway - is the right placing for Product Managers.