How AI is eating the Software World and why we dropped the Mobius loop from the Data-Driven AI logo

#humanwritten I was fortunate enough to be given a ZX Spectrum 48k back in the late 80s, and had an older brother (James Joyce the IT Tech bro, not the author) who got us into BASIC coding. At 19, I started working at Mercedes-Benz of South Africa as a Visual Basic coder back in the late 90s (yip - no university for me). I was in London picking up .NET during the internet boom (and bust) and had way too much fun in my twenties to appreciate or capitalise on the internet revolution (despite a few side hustles and game releases for Microsoft Pocket PCs).

After immigrating to Sydney, I consulted as a Silverlight developer (think the Grandfather of PowerApps) and then did a few long stints as Software Development Manager and Technical Lead in various ISVs and SaaS companies. In my day, you were a Senior when you had 20 years of experience and could write 5000-line SQL Stored Procs (I kid you not). Scrum came along, and everyone was suddenly "Agile". Software was usually client-server and consisted of a relational database, web server, API and MVC front end. iPhones changed the game again (I was a hybrid app developer in my spare time), and then Single Page Applications changed the game yet again and made the UX enjoyable for the user. 

The point is, before founding Data-Driven AI back in 2019, I was a battle-hardened software engineer perfectionist. I was that guy who made my dev teams read "Clean Code" and TDD books, and we tried (and failed) to implement Six Sigma in our software development life cycle. I realised back then that software was never really finished, it was constantly evolving until it was deprecated by the next big thing (usually when your competitors offered a better product and customer attrition started).

Why the history lesson, and how does this fit in with our new logo? It's a story about how software USED to be made.

The year was 2019. Data was still called Big Data, and Databricks was changing the rules of the cloud data game. Machine Learning at scale was becoming possible and popular thanks to Spark. Organisations were battling with their expensive, provisioned Data warehouses with silos and long time to value, and Data Lakehouses and the medallion architecture were becoming all the rage. It was the perfect time to start an AI-first company, and Data-Driven AI was born. Our vision was to help customers extract more value out of their data using AI, on Microsoft Azure.

We built one of the first enterprise Databricks Lakehouse platforms in Transport for NSW back in 2019 that allowed querying at scale of operational, multi-modal transport data with self-service analytics and reporting to make data-driven decisions and SLA efficiency analytics. This was the moment we knew the traditional Data warehouse architecture was dead. We had made the shift from application development to data and AI. (Side note: The best data engineers have a software engineer background and apply concepts like abstraction, DRY, etc, to their notebooks and pipelines. Show me a spaghetti notebook and I'll show you the door).

It was for this reason that we chose to put a Mobius strip/infinity loop in the Data-Driven AI logo to represent the continuous evolution of software whilst delivering data-driven and AI solutions. It should be pointed out that the AI we were building back in 2019 was Machine Learning and MLOps, not GenAI, however, I founded the company to help customers adopt ML to get more value out of all that data they had lying around.

Here's what our first logo looked like 6 years ago (thank goodness we didn't choose "Data-Driven ML" back then!). The more astute amongst you will notice that the Möbius loop is the 2 "d"s from data-driven. Mind-blowing stuff.

Fast forward 6 years, and AI no longer means ML. The market has rebranded AI to be shorthand for GenAI. Everyone has an opinion, of course, but the changes we are seeing to Software Engineering as a discipline are revolutionary. I've always said that the only reason we have coding languages is because machines cannot speak English. Well, now they can (and every other language). The software market is being shaken up and disrupted faster than any other market, because techies LOVE to try new things, and the first adopters will be the software engineers. Add to this the logic and structure of code, and it's ridiculously easy for AI to read and understand as compared to my 5-year school drawings.

The ability to WRITE software is no longer a primary requirement to BUILD software! This means your mother can have a good business idea, find the right AI tool, describe the OUTCOME they want, and AI will take you a % of the way. What this % is and how good the outcome is now, is not the point of this. It's X% of the way. And in 1 year it will be exponentially further than X% and so on until one day old-school Software Engineers, as I describe above, will go the way of the Data Scientist (RIP). 

"AI" is not a finished product, nor does it have a definition of where it ENDS. It doesn't end... it just keeps building on itself exponentially. Oh wait, maybe we should keep the Mobius loop after all!? No... It's time for a change.

We also noticed a disruption to the traditional consulting model. The old bait-and-switch of the Big 4, where an army of grads are assigned to a project to fight it out to rise the ranks to make Partner, are disappearing. AI is doing a lot of the stuff that the Grads would do in IT (I'm saving this for another post!). So now consulting firms are all about Quality, not Quantity, which is an awesome outcome for the customer. Your team should not be learning how to do consulting on customer projects!

AI has enabled small businesses like Data-Driven AI to punch well above their weight. We have productionised over 8 Enterprise Fabric data platforms in various state and federal government agencies, enabling large-scale analytics, AI-readiness and some amazing business use-cases. We have an agent in the app store. We have the world's number 1 top downloaded Azure FinOps tool in the Azure marketplace. But did you know that we are under 20 people? AI has enabled us to streamline our internal operations and processes, which results in higher profits, which means we can focus on finding and retaining highly-skilled consultants who love working on cutting-edge AI solutions and provide great outcomes for customers. It's win-win-win. I always thought we would organically scale to around 50 consultants, because that is what you were supposed to do when you built a business 5 years ago. Now I am questioning that model completely. We will scale AI internally and pair it with smart software builder consultants who can harness it for good. We might not ever go above 20 people if we don't need to. But let's get back to the logo refresh!

As the wonderful Kelly Wilkes said to me yesterday - "Scale up, Niche down, or Get Out". It's time for us to niche down yet again. For this reason, we are dropping the old school, clichéd Mobius Loop in our logo and emphasising the AI part of our name as we move into this exciting world as an established, AI-first Frontier Firm.

We only do data and AI on Azure, and we are the experts and have been doing it for 6 years with many amazing Microsoft case studies to back it up. Here's our new logo:

#azure #data #ai #fabric #microsoft #mspartner #frontierfirm #agenticAI