My Story · Julia Druck · Women Love AI

Who I am

Turning complex ideas into things people actually use

Before AI, I spent over ten years as a senior product manager in fintech, insurance and other regulated industries, leading product and the commercial side around it. I was product lead for AXA's motor insurance brand, Moja, and the customer journeys behind it. I even shipped an AI-powered advice product back in 2017, so this is not new ground for me. And before all of that, I started my career in the Royal Navy.

The through line is simple. I take something complex and make it work for real people and real businesses. That is exactly what I do with AI today, as a certified AI agent developer and trainer.

Some of the organisations I have worked with and for

The spark

I started small, and got hooked

I am not an engineer. I came at AI from the business side, so I started small: playing with low-code and no-code tools, doing a few bootcamps, building little things to see what was possible. Fifteen agents on MindStudio, automation flows in n8n, quick prototypes in Replit and Lovable. It hooked me fast.

MindStudio n8n Lindy Replit Lovable v0 Bootcamps, free and paid

The realisation

I could see what was coming

The more I built, the clearer one thing became. This was not a passing trend. AI was going to change how businesses work, all of them, and most were nowhere near ready for it.

And it was bigger than work. I have a son, Alex, growing up in a world this will reshape completely. I wanted him to have a head start with it, real access to the skills, maybe even a route into a job one day if that is what he needs. I could not give him that while standing on the sidelines myself.

Once I had seen all of this, I could not unsee it.

The leap

So I left, and started Serpin

So I made the call. I left my contract role and started Serpin with my husband Scott. It was a real leap, walking away from the safety of a contract to build something of our own. It is the best decision I have made.

Going all in

The way to get good is to never stop learning

Then I went deeper, and I have not stopped since. Free courses, paid courses, scanning what is new every morning with my own AI Radar. But the real lesson is simpler than any course: getting good at this is not something you finish. It is continuous learning, and the only way to keep it up is by doing, not by reading.

For me that happens two ways. I learn by building real things. And I learn by helping other people, through training and showing them how to use AI for themselves. Teaching is the part that surprised me: to explain something clearly you have to understand it properly first, so every time I train someone I get stronger at it myself. It is a skill in its own right.

What I've built

From no-code to Claude Code

I started with no-code and prototyping, then moved to building properly with Claude Code, where I could design real agentic architecture. There are two sides to how I put it to work: automating my own business, and solving problems for other businesses. These are just some examples, not the whole list.

Automating my own business

My second brain

Claude Code and Obsidian running my whole consultancy, with over 100 custom skills.

The Morning Brief

An autonomous agent that briefs me daily, running in the cloud.

AI Radar

A daily scan that keeps me on top of what is new in AI.

Solving problems for other businesses

Here the order matters. We start with the problem, never the tool. Only once we know it is real and worth solving do we build for it. A few examples of the kinds of problems this works well for:

Lead qualification

Sorting and scoring inbound enquiries so the right ones reach a person fast.

Tender responses

Drafting and assembling first-pass answers to long, repetitive tender documents.

Documents into process maps

ByLevel, our own product, turns business documents into clear process maps.

AI governance

A co-pilot that classifies AI systems under the EU AI Act.

Examples to show the shape of the work, not real client names. ByLevel is our own product. Confirm, trim, or reorder any of it.

How I decide what to build

Whether it is my business or someone else's, the same question comes first: what is worth automating, and what should stay human. Three questions decide it for me.

1Does it repeat? Predictable, happens often, the same shape each time. Good candidate.

2Does it need me? My judgement, my voice, or a relationship. If so, it stays with me.

3Is it draining? Eating time I should spend on the work only I can do. Worth handing over.

Automate the repetitive and draining. Keep the judgement.

Serpin today

What we do now

All of it rolls up into Serpin, the consultancy I run with Scott. We help businesses find where AI actually helps them, before anyone builds anything. Then we build it properly, with governance and quality built in. And because the hardest part is usually the people, not the technology, we train teams to use AI with confidence.

Find where AI helps

We start with discovery, mapping where AI fits your business and where it does not, so the effort goes where it pays off.

Build it properly

Agents and workflows with governance and quality built in, not demos that fall over in production.

Train your team

Hands-on training that makes AI accessible to non-technical people, so the capability stays with you.

I built you one of my tools

The sales brief is one example of all of this. Have a play with it, and come and say hello on LinkedIn.

Build the sales brief Connect on LinkedIn

From using AI to building with it