What does it take to replace a system of record?
As a self-professed word nerd, I find the emergence of new words and terms fascinating. I'm old enough that in my first year in intermediate school, we had Latin as a subject. I distinctly remember my teacher talking about how Latin was a 'dead language' because there are no native speakers and it isn't used in everyday conversation. Another way to think about it is that it didn't die, per se, it just evolved into the Romance languages like Spanish and French. It also heavily influenced English too, of course.
But back to Latin. Because it is used as the official language of the Vatican, there are still new words created for the language to allow for modern terms to be referenced in modern documents. (Did you know the Latin for email is epistula electronica?)
Just the idea that language continues to evolve, organically, is interesting to me. It also makes me wonder what earlier speakers of English might think of some of our more modern vocabulary or tech slang.
Take "vibe coding" for example. Sure, ancient types might literally translate it to programmatio sententiâ, but it may as well be translated to "magic". I mean, programming in general is pretty amazing when you stop to think about it. And now we are at the stage where you don't even need to know coding syntax and you can just speak your intent into being?! As someone with a technical background it is both amazing and a bit alarming when you start to think about some of the implications.
Amazing seems to be winning out, though. There is plentiful commentary available talking about how vibe coding is going to transform the world of software as we know it. That anyone can now spin up a new system and change everything overnight. Or that you can vibe code a system of record in a weekend that will replace the existing systems of record just like that.
However, I don't think it is quite as simple as that.
Change is simple…said no one ever
I have always liked the definition of a system that evolved from Dr Harold Leavitt's work in organisational theory, was later popularised by Bruce Schneier, and was then studied by me when I was at university. The so-called "golden triangle" whereby we see a system as the combination of people, process and technology.
The reason why I keep coming back to this definition is because all through my career, whenever I was helping lead any kind of organisational change, the interplay between all three elements kept coming up. Even though, as a solution architect, the initial expectation was almost always to that I would care the most about the technology change, time and time again the harder change was the processes and the people.
I even remember discussing the nature of this interplay with a good friend of mine when we were at uni; she was studying psychology, and I was studying information systems. She remarked that her degree was really all about people, and I said "same." To say she was a bit surprised at my take would be an understatement. "No," she said, somewhat indignantly, "yours is all about technology." To which I responded, "nope — it's about people using technology to get things done."
People using technology
Look, it's pretty clear that advances in generative AI have fundamentally changed some of the 'factors of production' when it comes to writing code. Not for the first time, new tools are making it easier than ever before to bridge the gap between the business or customer need and the tech providing capability to help meet that need. Just expressing that need in the right way can get some pretty impressive results, and I say this coming from my own experiences with these capabilities.
I've long pointed out that AI is really about augmenting our intelligence. And as someone with a background in software development, these latest tools have really supercharged my ability to get things done. So, I am far from a detractor when it comes to the potential of these tools.
But when talking about 'systems of record' or any type of system, really, I keep coming back to the fact there is more to those than just the tech. Much more. In fact, that "golden triangle" is rarely an equilateral one, in my experience. The three components don't usually carry equal weight in every system. And even if you can find a system where they do, I think that the two corners of people and process would usually combine to outsize the tech in terms of effort required to affect really meaningful and lasting change.
ADKAR
So, when it comes to managing change in the non-technology parts of the system, it's probably no surprise that I have a preferred framework for helping with that activity. The ADKAR model by Prosci is something I was introduced to many years ago, and it has stayed in my toolbox because of its simplicity and memorability. It has also been really useful as a reminder to think about multiple dimensions of change.
There is a whole book you can read on the framework, but I'll summarise my take on the components of ADKAR:
- Awareness of the need for change and why the change is being asked for or encouraged.
- Desire to take part and be supportive of the change(s).
- Knowledge of what it will take to affect the change
- Ability to make the changes required to competencies and processes
- Reinforcement of the change so that the system doesn't just revert to the old way of doing things.
As I say, the memorability of the framework has meant that I keep coming back to it and using it as a tool to challenge my thinking and check my biases. In planning a change, or when seeing resistance to change, ADKAR has prompted me to ask questions that I might have otherwise skipped over: have I made it clear enough what the change is and why it is important? Can I see where there is a blocker to getting a team to buy in to the change? Have I provided enough training and coaching on the new skills required in order to change? Have I effectively changed the underlying process in the right way to enable the change?
Very useful.
The scripted 'R'
I used to joke that I would often see change plans that used ADKAR the way that many Kiwis in the North Island would pronounce it — i.e. without the 'R' at the end. Yet, the reinforcement of change is often the hardest and most critical piece, in my experience. This is just one excellent example of where humans and machines differ!
With the tech, once you write, test, compile and ship the code, the change in that part of the system is implemented (for the most part). But humans have memories, preferences, habits and autonomy. Creating the environment and the incentives to embed and reinforce change is often the hardest part — especially when that part needs to keep working even after the 'change project' is all wrapped up.
Think of the last time you consciously tried to change an established personal habit or practice. How hard was it to truly embed a new way of doing things? A personal example, speaking of the letter 'R': I once decided that I wanted to start to handwrite using the cursive 'r'. I was never taught how to do so when I was a child learning how to write, so this was a whole new thing for me. The problem is that, even after deciding to make this change that I wanted to implement (Awareness + Desire), and even after looking into and practicing the techniques for writing the 'r' in this way (Knowledge + Ability), I kept falling back to writing the way I always have.
Muscle memory is a hard thing to change, and it is even harder when I don't actually write by hand that often anymore. So the traditional ways of learning and reinforcing through regular repetition aren't tactics that will work anymore. I think I'm finally getting the hang of it now, but that has taken much longer than I ever expected at the start.
You can see where I'm going with this, I hope? Changing systems requires changing the technology and people and the process. And some of those changes take far longer than buying a nice new ballpoint pen.
Can you vibe code a replacement system of record in a weekend? Maybe the next time someone breathlessly claims that, ask about the processes that would need to change to adopt the new tech and how much planning is going into changing those. And then chat about the people who are more than users of the system — they are components of the system themselves. How will they need to adapt and change to adopt and leverage the new tech?
Can all of that happen in a weekend? Maybe. But unlikely.
