By Josefine Boqvist – Founder of Deerdata
I’m noticing a pattern play out with AI.
For years, the mantra was that data is a strategic asset. Most organisations agreed in principle – and then stored it in multiple different places.
Every role in software development is getting faster. Business analysts, designers, developers, testers – everyone has a co-pilot now. And yet I keep hearing the same thing from teams:
“I’ve explained the same thing to the AI a hundred times.”
Sound familiar?
The problem isn’t the tools. It’s that we’ve optimised for individual speed while leaving organisational knowledge exactly where it’s always been – fragmented, inaccessible, and walking out the door every time someone changes jobs.
Where the knowledge actually lives
Requirements are kept in in one tool, and design decisions in another. Architecture choices are stored in a PowerPoint deck from a few months ago. The reasoning behind a technical decision is buried in a Teams thread.
AI has made people faster at working inside this system. It hasn’t fixed the system.
Three levels of maturity
I think about maturity in terms of three levels.
Level 1 – AI in Silos: Individuals use AI to do their work faster. This leads to real gains, but knowledge stays trapped with the person.
Level 2 – Shared Project Context: Documentation, decisions and discussions live in a shared layer that AI can access. Now the whole project benefits, not just the person at the keyboard. This requires deliberate effort – decision logs, living documents, records of what was changed.
Level 3 – Company Memory: Every project generates knowledge worth keeping: integration patterns, security decisions, lessons learned the hard way. The problem is that most of it disappears. The organisations building real advantage are treating this knowledge the same way the best ones treat data – as something worth capturing, structuring and building on.
Projects consume and create context. Over time, the need for a structured knowledge base compounds.
The shift worth making
AI-driven coding makes developers faster, while AI-driven software development makes organisations smarter.
As code gets cheaper to produce, context becomes the scarce resource. The companies that thrive won’t be the ones generating the most code. They’ll be the ones whose AI gets better with every project – because they invested in the knowledge layer underneath it.
Most of us are still optimising the wrong thing. I was too, for a while.
The shift from individual productivity to organisational memory – that’s where I think the real leverage is hiding.
—
If you want to learn more about where your organisation sits on the three maturity levels, reach out at info@deerdata.se.
