Today(ish) is more or less 30 years since I founded Codify. We’ve been through a few iterations over the years, and I thought I would put down a few emdash and emoji-free thoughts.
We started out helping customers by writing custom software – SQL Server 6.5 + .idc/.htx scripts on IIS on NT3.51 (pre-dating Classic ASP … really showing my age) – today we’re a leading Microsoft Azure Cloud Solution Provider helping a variety of amazing customers get the most value out of their investment in Microsoft’s Azure Cloud Platform.
We wrote a number of fairly sizable systems for customers in our early days. At that time the word DevOps didn’t exist and the sorts of software engineering methodology tools that are free now were $20K a seat back then. Most of the development tooling for the web (Visual InterDev anyone?) was often more of a hindrance than help and VI was always a good option. Source code management tools were stone age (SourceSafe) and prone to corruption. Source code on a fileshare was commonplace due to the cost of the commercial options.
I saw the writing on the wall in 2010 – it was clear to me that the future would be SaaS or low-code platforms based on an xrm like Dynamics. I saw this as an inevitability and so I opted to reinvent the company around Microsoft Enterprise infrastructure stack in 2011 (this is when Ben joined the company as Director). While I had not really played in this space before – this is where Ben grew up. Bespoke software engineering is a difficult and high-risk endeavour (especially with the tooling back then). In looking at the numbers it was clear that shorter, more precise MS infrastructure deployments were easier to deliver, the value easier to articulate to customers, and the overall project delivery overhead much lower. In retrospect I might have made this call a few years too early however it has become true that there is a lot less bespoke software development these days. I didn’t really anticipate how far the industry would come with better tooling though. No one saw agentic coming.
In 2014/15, we ended up in a sizeable Cloud infrastructure project with Microsoft Services delivering the Cloud Services Broker for Queensland Department of Education. Azure was still quite nascent at this time and to be honest, the initial infrastructure offerings (Azure Cloud Web/Worker roles) were challenging to work with. The original Azure was built on an assumption people were going to write apps in terms of web/worker roles – that was never going to happen – and never did. As that project wrapped up, Microsoft released the Azure we know today – Azure Resource Manager, RBAC, and most importantly, Infrastructure-as-a-Service. It turned out most enterprises just wanted VMs.
We continued a lot of our Enterprise infrastructure work however watching the rapid adoption of Exchange Online, the maturation of Azure AD and associated technologies, SharePoint Online, etc it was again obvious to me that the entire Microsoft Enterprise approach was going to go under the bus and it was time to change again.
We began providing Managed Cloud Services to customers shortly after and needed to quickly come up to speed on how to manage a business that needed to move from Professional Services (big, intermittent sugar hits of revenue) as opposed to Managed Services (smaller, incremental, but importantly, accretive hits of revenue). We were able to lean on years and years of deep experience in software engineering to help customers understand the work required to get custom apps from on-prem to the cloud. Our deep experience in the Microsoft Enterprise stack through domain consolidation projects and the like meant that we could help people build infrastructure that could straddle both the on-prem and cloud worlds at the same time.
Since starting on the CSP path we’ve been recognised by Microsoft as Tier 1 / Direct CSP status specialising in Digital & App Innovation and Infrastructure – both just the cloud versions of where we started out in two previous reboots.
Looking forward to our next reinvention, AI does not faze me. It is another evolution of IT – a tool like any other (albeit the most amazing one I’ve seen since starting out in IT in the early 90s).
A lot of people are sensitive about AI putting people out of jobs. The harsh reality is that it is not going away – the genie is not going to be put back in the bottle – and you must face it as an opportunity to reinvent what you do – it is not something to be challenged or objected to.
There are two constants in information technology – change and commoditisation. The only people who AI will put out of work are those who choose not to adapt. It is your choice to make; as I am fond of saying in company meetings, AI is going to put a lot of people out of work, and we better make sure we’re helping our customers be the putters and not the puttees. If they are successful, then we are too.
I’ll close out this note on a bit of a sombre reflection, and that is the increasing bureaucratisation of business in general. I started Codify when I was 23 on the sniff of an oily rag. Today, to get a competing business off the ground, the barriers to entry are simply too high for young entrepreneurs to engage in mid-market. The entry level today is PI/PL insurance, panel memberships everywhere, ISO 27001, ISO 9001, endless audits and certifications. This is all, in the scheme of my career, a fairly recent invention. We’re completely cutting out the next poorly funded start up that might grow into an enduring and successful business that leaves indelible improvements on its customers’ systems and processes. This is something that disappoints me. By systematically making it harder for new entrants we are lowering customer choice.
We’re doing the same thing to software engineering with AI and I’ll write about that in a future post. The real challenges with AI are going to be around hollowing out of expertise – and this won’t manifest for a few years.
David Connors
Founder and Managing Director, Codify