The question isn’t whether you are ready for AI.
It’s whether your data is.
Because some organizations have been recording their history for years — and others are still only keeping snapshots.
The difference doesn’t show much today, but it soon will.
Where AI meets Event Sourcing
Last week I attended the AXONIQ Conference in Amsterdam, where the main theme was “Where AI meets Event Sourcing.”
It wasn’t my first time at that conference — but this time I joined literal round-table discussions. Architects, C-level leaders, and AI specialists sitting together, finally having the kind of conversations that felt overdue.
Those discussions were eye-opening — not because of any new revelation about AI itself, but because it became clear what Event Sourcing has quietly been preparing us for all along.
Decades of recording intent, state changes, and causality — and only now the world is catching up to why that matters.
The gap is already here
I’m not an AI expert.
But I am an Event Sourcing practitioner — I’ve built, released and maintained for years two event-sourced systems, one in 2012 and another in 2016.
Back then it felt niche, almost academic.
Today, it suddenly feels… strategic.
Here’s the uncomfortable truth:
They’ve been training, in slow motion, for the moment that just arrived.
And the gap between these two worlds is compounding every single day.
Every day you don’t store your events, you lose a day of potential learning.
Every decision or correction made by your users that disappears from your database is one less data point your future models will ever see.
A flat picture or a timeline?
I once wrote a post titled “Event Sourcing is like vector graphics”.
That was in 2012 — thirteen years ago — and I’m not sure I’d put it that way again today.
But back then, it made sense: I was looking for metaphors to explain that Event Sourcing isn’t about storing state, it’s about keeping the whole path that led there.
The idea still holds.
CRUD systems store pictures.
Event-sourced systems store timelines.
And when AI arrives, it’s the timelines that will matter — because they contain not just what happened, but how and why.
Why history matters for learning
by orders of magnitude more."
That’s not a slogan — it’s reality.
AI models can learn patterns of change. They can learn why things evolved, not just what they ended up as.
Snapshots tell you what your business looked like yesterday.
Events tell you how it got there.
It’s easy to underestimate how quickly this difference will become visible.
When companies start training their in-house LLMs — and many already are — those with historical event data will get richer, more accurate models from day one.
Others will feed their AI a single static state — like teaching a historian by showing them only today’s newspaper.
The spectrum of readiness
There will be companies starting their Event Sourcing journey tomorrow.
Others will start in a year, or five.
And some already have years of rich domain history quietly stored in their event logs.
The question is:
When your company decides to train its first serious AI model — on your own data, for your own purposes — what will it find?
History or just a snapshot?
Because by then, it will already be too late to go back and collect that history.
All you can do is narrow the gap.
PS — If you want to start narrowing that gap
If you want to build that capability — to store not only data but the evolution of your business — that’s what I teach.
I run hands-on trainings on Event Sourcing with Axon Framework:
patternapplied.com/en/trainings/cqrs-event-sourcing-axon
I’m also (slowly 😅) building an online course, CQRS Applied, where I explain how to use Event Sourcing and Axon Framework in Java step by step — realistically, and without hype.
You can download my free e-book (in Polish) “4 Tools for DDD/CQRS/ES on the Java Platform” there and join the mailing list.
— about a training, consulting, or just exchanging ideas
— reach out:
And if you’re looking for other ways to work together — reach out.
We’ll figure it out. 😉
I always respond.
Comments
Post a Comment