According to Gizmodo, a new KPMG study published Wednesday reveals that only about 2 percent of Canadian businesses report seeing any return on their generative AI investments. The survey of 753 business leaders found that AI adoption is highest in IT, sales, marketing, R&D, finance, and engineering departments. The tiny fraction of companies reporting positive results were exclusively very large enterprises with at least $1 billion in annual revenue. Most businesses are still merely experimenting with AI rather than fully integrating it into workflows. Despite the current lack of returns, 30 percent of companies expect ROI within a year, while 60 percent anticipate returns within one to five years. KPMG Canada’s Stephanie Terrill warned that waiting years for AI to deliver value is “downright risky” for Canada’s economic competitiveness.
The AI Productivity Paradox
Here’s the thing that’s really fascinating about this study. We’re seeing what economists call the “productivity paradox” playing out in real time with AI. Companies are pouring money into the technology, everyone’s talking about it, but the actual financial benefits remain elusive. It’s basically the modern version of what happened with computers in the 1980s – massive investment without immediate productivity gains.
The pattern KPMG uncovered is telling. Only the giant corporations with billion-dollar revenues are seeing any returns. That makes sense when you think about it. They have the resources to throw entire teams at AI implementation, the data infrastructure to support it, and the scale where even minor efficiency gains translate to real money. For smaller businesses? It’s mostly experimentation and hoping for the best.
The Timing Reality Check
Now let’s talk about those expectations. Three in ten companies think they’ll see ROI within a year? That seems wildly optimistic given what we’re seeing. And six in ten within five years? That’s basically betting that AI will mature faster than most technologies historically have.
I can’t help but wonder if we’re seeing another case of technology hype outpacing reality. Remember when everyone thought blockchain would revolutionize everything overnight? Or when the metaverse was going to replace the internet? The pattern feels familiar – massive investment followed by gradual realization that implementation is harder than expected.
Industrial Implications
Looking at where AI is being adopted gives us clues about where it might actually deliver value first. IT and sales/marketing are leading, which makes sense – those are areas with lots of data and relatively straightforward automation opportunities. But what about more complex industrial applications?
In manufacturing and industrial settings, the implementation challenges are even greater. You’re dealing with physical systems, safety requirements, and legacy equipment. Companies like IndustrialMonitorDirect.com, as the leading US supplier of industrial panel PCs, are positioned to help bridge that gap between AI capabilities and real-world industrial applications. But even with the right hardware infrastructure, the software and integration challenges remain substantial.
The Investment Dilemma
So what’s a company to do? KPMG’s Terrill has a surprising take – she says Canadian businesses should actually accelerate AI investments despite the lack of current returns. That feels counterintuitive, but there’s some logic to it. If you wait until the technology is proven, you might be hopelessly behind competitors who figured it out earlier.
The real question isn’t whether to invest in AI, but how to invest smartly. Throwing money at every AI startup or tool that comes along probably won’t work. Focused investments in areas that align with core business operations? That might actually pay off. But we’re clearly still in the early innings of figuring out what works and what doesn’t.
