Jumping straight into innovation. What a waste of time.

Life is a set of pyramids. With pyramids, each layer depends on the layer below it, and if the layer below it is not solid, everything on top is at risk of falling down. And if it's really not solid or even not present, then it's definitely going to fall down.

Let me give an example to move it from an abstract metaphor into something a little more real.

The testing pyramid is unit testing, then integration testing, then UX testing, and finally manual testing. If any of the lower layers are not solid below, the layers on top are going to collapse. The reason why almost all testing of most applications goes straight to manual testing which is a nightmare and takes months is because the unit testing, integration testing, UX testing, and automated UX testing are not in place. Similarly, the deploying of that infrastructure, the monitoring around that infrastructure, the deployment of an application, and then the monitoring of that application is a set of steps. If each of those layers is not in place, the whole thing comes down.

You've got to get your foundations in place first.

The reason I think that so many enterprises are struggling and they try to solve it by “doing innovation” and “bringing in the smartest people in the world” and why it always fails is because they jump to the top layers of the stack. They ignore the fact that actually, you've got to build really solid foundations, (which is really hard and really boring) before you can start doing the next layers up.

Jumping straight into innovation is such a waste of time. How many companies right now are talking about AI? “We're gonna bring in generative AI, we're gonna build an AI workflow.” You guys still have servers sitting under someone's desk that goes down once every 24 hours because Bob kicks the cable out! What are you talking about AI? Forget that, you're wasting your time jumping to higher-order levels of innovation when your foundations aren't in place.

If I was the CIO of a bank, I would scan around, and I would look for the foundational pieces that needed fixing. And I would put all my effort into that and everything else would be ignored. Business users would hate it, but they'd have to trust me. I'd have to have a CEO which gave me that sort of strength because it's not like, “oh, I'll do it in parallel.”

What's amazing is, if you talk to your ops people or your tech ops people, they probably already know the answer - they've probably been shouting and screaming for the last 10 years. Yet we'd go and get an expensive management consultant to come in who will interview everyone, build a nice slide deck to then be presented out, and ‘wow’ management. But when you show it to the ops person, they tell you, “I could have told you that. In fact, not only could I have told you, I did tell you! The management consultant’s just telling you what I've already said!” These foundational pieces, if they're not in place, nothing else is.

When I was working for Amazon, we would go in and talk to CIOs and heads of engineering and show them our cloud stack and the solutions we could provide. And they'd be like, “I heard that Amazon's got some really cool AI, can you show me your AI models?” Or “I heard that you can do this virtual reality sales booth so I don't need bank branches anymore?”

Why are we talking about AI and VR bank booths when you don't have the absolute fundamentals of your infrastructure in place? It still takes six months for someone to get a server, where they have to predict what their server usage next year is gonna be today. And you're talking about AI and big data?

Come on. You’ve gotta get your infrastructure in place.