Transformations
The Past
Let’s take a trip through time … Step into the Tardis and let’s go!
Agile and Scrum

Agile, and its most famous implementation Scrum, were a response to the overly bureaucratic software engineering of the last century. We all remember UML and 100+ pages documents full of specifications for each variable name. Agile, and Scrum, promised x10 more productivity, fast iterations, and software in days rather than years.
Did it deliver?
Kind of … Many companies continued doing what they were always doing, paid lip service to Agile/Scrum, and found no benefits. This was unsurprising. However, many companies did implement Agile/Scrum and found that all was not as promised. Yes, a lot of things got better. But it was not a x10.
Virtual Machines

I remember the first talk from VMWare which was done on a Linux machine running on a Windows machine which was running on a Linux host. It was epic. It was going to revolutionise all hosting. No more servers, just virtual machines all over the shop running on a few super massive hosts. All system administrators will be out of work, as you needed just one.
Did it deliver?
Kind of … There were still images to tweak, update, and backups. Host machines still needed patch cycles, hardware maintenance, and all the VMs were fighting for the same resources. More monitoring, more work, more patching. There were great achievements here, the ability to run software on old versions of OSes on demand, hosting of personal or small business machines at low costs. But, it was not a x10.
Docker

Docker came along because managing virtual machines was a pain. Let’s have a thing that will orchestrate your operating system so it runs the minimal amount to run your service. Perfect! No more configuring images, paying licences fees, and having to upgrade all the time. It was just one command. Of course, the real benefit was to stop developers saying “Well, it works on my machine”. Fine, we’ll ship your machine then. It was going to be fast, easy to use, and give us x10 more productivity.
Did it deliver?
Kind of … It was hard to use, the specifications changed all the time, and there was no
orchestration between docker images. Yes, docker compose came along. Running that in
production is problematic to say the least. Unless you’re happy with 90% uptime. Base
images always contain something you don’t want, they lack prompt security updates, and
some even have bash and ssh installed by default. Still, everyone now uses them
because they are so useful. But, it was not a x10.
The Cloud and Kubernetes (K8s)

Let’s move to the Cloud. No more system administrators, no more machine rooms, no more air-con failing, no more hassle. Let someone else worry about it all, we just use it. It’s going to save us x10 the costs! Look at how cheap it is, just $0.00002 a minute. Best of all, we can have the data/services in different regions for fast access. x10 more productivity, here we go.
Did it deliver?
Kind of … The cloud is just someone else’s computer. There’s no security there. It’s massively complex so all your sys admins that you thought you’d get rid of are now Platform Engineers, or SRE. Their super power is to transform YAML into AWS bills. Of course, Cloud is super useful. But, it was not a x10.
The Present
Agentic Transformation

Let’s use AI agents for everything. x10 productivity boost guaranteed. No one will ever know it’s all AI, from vibe coding to LinkedIn outreach, we can make it so our Claude skills teach the AI not to use em-dash — or whatever you chose. Note that this was💯 written by a human, no LLM here. Maybe an image or two.
Will it deliver?
Kind of … AI is an amplifier therefore, it amplifies the good to make good faster, the okay to bad, and the bad to worse. Chances are, your dev processes are nowhere near good enough to be amplified. And if they are, what happens to the rest of your pipeline?
The AI field changes monthly, and no one knows what the best way to do things™ is yet. Anyone who tells you otherwise is probably selling something. Yet, we do know what does not work. And that is valuable. Between safety, scalability, and economics of AI, where do you stand?
Will it be a x10 or x100 boost or a massive negative? Well, the Forbes Research 2025 AI Survey found that less than 1% of the 1,075 C-suite members polled had seen a significant ROI, defined as a 20% or more increase in profitability or cost savings. Only 3% report a substantial ROI of 10-20%, while the majority (53%) reported a limited ROI of 1-5%.
The PWC’s 2026 CEO Survey states that 12% of CEOs say AI has delivered both cost and revenue benefits, while companies that have scaled AI with strong foundations are pulling ahead.
The future does not look as bright as the AI evangelists would like us to believe.
Conclusion
How would you like to be in the 12% that show both cost and revenue benefits?

We are working on an Agentic Transformation framework. It’ll be released early next year. If you’re interested, please do reach out to us and you’ll be the first to know. As a tiny spoiler, here’s a preview:
- Get organisations to treat agents as governed production systems.
- How to get deep domain knowledge humans controlling AI agents efficiently.
- Control the cost of tokens by using AI safely and intelligently.
- Know how AI fails, before the board, your customers, or a regulator finds out.
- How to enable scaling of AI software without the headaches.
Why? To enable ideas to move from business concept to software faster.
…
Imagine, if you will, a day when all the drama is removed from your software production: no panic, no crisis, just smooth software releases that exceed your customer’s expectations. This is what we have done in the past and can do for you.