What the 5 per cent of companies getting real value from AI understand

Management

Sponsored by Mindstone

Joshua Wohle, CEO, Mindstone

The gap between AI’s promise and its payoff has become corporate technology’s most expensive problem. Gartner projects global spending on generative AI will reach $644 billion in 2025, yet the transformative productivity gains remain stubbornly elusive.

For Joshua Wöhle, CEO of employee AI upskilling platform Mindstone, the diagnosis is simpler than many executives want to hear: the problem isn’t the technology. It’s how we’re preparing people to use it.

“Fewer than 5 per cent of organisations are really getting value from AI right now,” says Wöhle. It’s a stark assessment, drawn from Mindstone’s work training teams at Fortune 500 enterprises, though one that echoes broader industry findings. BCG’s October 2024 report “Where’s the Value in AI?”, which surveyed 1,000 executives across 59 countries, found that 74 per cent of companies have yet to show tangible value from their AI investments.

The implication is uncomfortable: most organisations aren’t failing because they chose the wrong tools. They’re failing because they’re asking the wrong questions about how their people should work alongside AI.

Mindstone, which delivers AI skills training for enterprises, has built its approach around this diagnosis. Asked how the company closes the gap, Wöhle points to two key components. First, overcoming what he calls “suspension of disbelief”: the lingering scepticism many professionals feel towards AI. “We have not found anything that works better than a live demo,” he says. Second, making AI directly applicable to each employee’s day-to-day responsibilities.

Augmentation over automation

The crux of Wöhle’s argument rests on a distinction that sounds semantic but proves profound in practice: automation versus augmentation. For decades, technology investment has been synonymous with automation, replacing human effort with machine efficiency. Assembly lines, spreadsheets, CRM systems: all designed to do work humans used to do, only faster and cheaper.

AI doesn’t fit neatly into that paradigm. When organisations treat generative AI as another automation tool, they set themselves up for disappointment. “In automation, AI actually is overhyped,” Wöhle admits. “It can make a lot of mistakes.” The technology hallucinates, misses context, produces plausible but incorrect outputs. It’s a failure mode familiar to anyone who has encountered AI-generated statistics that sound authoritative but are made up.

But that same technology becomes transformative when deployed differently. Augmentation, using AI to enhance rather than replace human capability, is where the real value lies. An automated system attempts to eliminate the human from the loop. An augmented system keeps the human firmly in control, with AI serving as an amplifier of judgment, creativity and expertise.

This distinction explains why so many AI deployments disappoint. Companies buy sophisticated tools expecting them to work autonomously, then grow frustrated when outputs require constant human correction. They’ve purchased automation and received something that demands augmentation. The mismatch breeds cynicism and wasted budgets.

Why mindset matters more than mechanics

If the technology itself is capable, what’s holding organisations back? According to Wöhle, the answer is unexpectedly human. “Only about 10 per cent of what we do is helping people use the tools,” he explains. “90 per cent of this technology is about a mindset shift.”

This flies in the face of conventional training programs, which typically focus on feature tutorials and workflow documentation. Click here, type this, export that. Such training assumes the tools are the barrier and that competence follows from familiarity. It’s the same approach organisations have used for software rollouts for decades, and it’s precisely wrong for AI.

The real barrier is conceptual: people need to fundamentally reimagine how they approach work in a world where AI can handle portions of it. This isn’t about learning keyboard shortcuts. It’s about learning to think differently about problems, outputs and what constitutes valuable human contribution.

Demonstrating value in real time

That first challenge, suspension of disbelief, runs deep. Many have been burned by previous technology promises. They’ve sat through demos of products that worked beautifully in controlled conditions and miserably in practice.

To break through this barrier, Mindstone relies heavily on live demonstrations tailored to specific roles. Abstract promises mean little to sceptical professionals. What shifts minds is witnessing AI solve their actual problems, in real time, with their actual work.

In conversation with Business Reporter, Wöhle demonstrated this principle in action. Rather than describing how AI could help with interview preparation, he walked through the process live: using voice input to create research briefs, feeding outputs between different AI tools and iteratively refining results. The demonstration took minutes but produced interview questions, background research and a structured script that would have taken hours to assemble manually.

The approach revealed something critical: effective AI use isn’t about finding the perfect prompt. It’s about providing rich context and using AI as a thinking partner rather than a vending machine. “When you speak to your computer,” Wöhle observes, “you intuitively provide it with dramatically more context.” This contrasts with search engine habits, where shorter queries yield better results. With generative AI, the opposite holds true.

The stakes of getting this wrong

Demonstrations break through scepticism, but lasting change requires connecting AI capabilities to specific job responsibilities. A marketing manager needs to see AI transforming their campaign planning. A financial analyst needs to see it accelerating their modelling. Abstract potential matters less than concrete relevance.

For organisations still treating AI adoption as a technology procurement problem, the risk isn’t just wasted software licences. It may be competitive ground ceded to rivals who understood earlier that the bottleneck was human, not technical. Wöhle believes the gap between AI’s early adopters and the rest is widening, and it’s not closing on its own.

His diagnosis offers both challenge and opportunity. Meaningful AI adoption requires investment in people, not just technology. But that investment doesn’t demand massive budgets. It demands a willingness to rethink how people are prepared to work alongside AI. The organisations that recognise this have a window to pull ahead. That window won’t stay open indefinitely.

Build AI capability across your organisation with Mindstone – book a demo today