Let’s be honest… too many organisations are still running on legacy systems held together with duct tape and hope. Slapping AI on top won’t magically fix poor fundamentals – it’s like bolting a massive spoiler and large exhaust onto a 30-year-old car with a rusting frame and a clapped-out engine. It might look impressive, but it’s still unreliable and inefficient.
The legacy tech problem
According to Mechanical Orchard, a staggering 60% of enterprises still rely on legacy systems, with some industries, like banking and accounting, running on technology that’s decades old. And that’s being generous…
In fact, another study by Forbes found that 70% of digital transformation projects fail due to outdated infrastructure. Businesses keen to adopt AI often overlook the fact that their existing tech stack isn’t ready to handle the demands of machine learning and automation. Investing in AI without fixing these core issues is a recipe for wasted time and money.
Rubbish in, rubbish out
AI thrives on clean, structured data and robust infrastructure. If your current systems are slow, disjointed, and riddled with inefficiencies, AI won’t save you – it will just amplify the mess. For example, a survey by Gartner found that poor data quality costs businesses an average of $12.9 million per year. If you’re feeding AI incomplete, inconsistent, or inaccurate data, don’t expect ground-breaking insights – expect more confusion and bad decisions at scale. Rubbish in, rubbish out.
Before you even think about AI, you need to modernise your core technology, streamline your processes, and get your data in order.
Otherwise, all you’re doing is automating inefficiency.
Solve a real problem, Not just a trend
But even before all this, you need to ask yourself one core question: do you actually need AI, or do you just want AI?
It’s easy to get caught up in the hype, but AI should be solving a real business problem. According to a McKinsey report, while AI adoption has increased, only 20% of companies see significant financial benefits. Why? Because many adopt AI without a clear use case, leading to expensive experiments with little return.
If you’re not clear on the ‘why,’ then you’re just adding complexity for the sake of it.
My top tip: focus on fixing the basics first – then, and only then, should AI (or any other ‘next step’ technology) enter the conversation.
Sort your fundamentals, clean up your tech stack, and then see if AI genuinely adds value. Otherwise, you’re just a teenager revving a rusty car with a loud exhaust – making noise but going nowhere fast.
