88% of companies already use AI in at least one function. But how many are truly transforming their processes? Here are the AI developments that matter for businesses in 2026.
Nearly nine out of ten companies already use artificial intelligence in at least one process. That’s the most striking figure from McKinsey’s latest State of AI report, which captures a landscape where AI is no longer an experiment for the few: it is a widespread reality, but one that is still largely unfinished.
The real question in 2026 is no longer “should we adopt AI?” It’s “why are the results slow to arrive?”
From Pilot to Scale: AI’s Bottleneck
The problem isn’t the technology. It’s the transformation.
Two-thirds of the companies surveyed by McKinsey are still in the experimentation or pilot phase. Only a third have begun scaling AI across the entire organization. Among large companies (with revenue over $5 billion), that share rises to nearly 50%, but for small and medium-sized businesses it remains stuck at 29%.
The most common bottleneck? Failing to redesign processes.
Companies achieving concrete results share one trait: they don’t simply bolt an AI tool onto existing workflows. They redesign those workflows from the ground up around the new capabilities. “High-performing” companies are nearly three times more likely to do this than others.
The Most Disruptive Development: AI Agents
The most significant trend of 2026 is the rise of AI agents — autonomous systems capable of planning and completing sequences of tasks without continuous human intervention.
62% of companies are already experimenting with AI agents in at least one function. 23% have scaled them across parts of the organization. The areas where they are spreading fastest are IT, knowledge management, and customer service.
What actually changes? An AI agent doesn’t just answer a question. It analyzes the situation, retrieves information from company systems, and proposes or carries out an action. That’s the difference between a passive assistant and an active collaborator.
For companies operating in industrial and field service settings, this has direct implications: it means having AI support that, during a technical intervention, accesses manuals, interprets machine data, and proposes a solution — in real time, directly within the operational environment.
Where AI Creates Value (and Where It Doesn’t Yet)
McKinsey’s report draws a clear distinction between perceived benefits and impact on the bottom line.
The most frequently cited benefits at the company level are improved innovation (64%), customer satisfaction (nearly 50%), and competitive differentiation. The impact on EBIT is less immediate: only 39% of companies attribute measurable financial benefits to AI.
The sectors showing the clearest cost advantages are software engineering, manufacturing, and IT. Revenue benefits, on the other hand, are concentrated in marketing and sales, product development, and strategy.
In other words: companies using AI to streamline production and reduce machine downtime are starting to see the numbers. Those using it only to generate content or answer emails, less so.
The Most Underrated Risk: Inaccuracy
While companies race to adopt new solutions, the main risk highlighted in the report is the inaccuracy of AI output.
Nearly a third of companies have already experienced negative consequences linked to AI errors. Yet many still haven’t established clear processes for determining when and how to validate model-generated results.
For those using AI in critical operational contexts — maintenance, technical support, plant management — this isn’t a minor detail. It’s a prerequisite: AI must be built into systems that guarantee accuracy, traceability, and human oversight at the moments that matter.
What This Means for Industrial Companies
If you work in a sector where maintenance, field service, and operational continuity have a direct impact on results, the message is clear.
AI isn’t useful only in the office. It’s useful in the field.
Technologies that combine remote assistance, extended reality, and AI integrated with corporate documentation systems are redefining how field technicians work: procedures accessible during an intervention, remote expert support, instructions contextualized to the task at hand. Fewer errors, less downtime, fewer site visits.
This is exactly the model we are building with XRSupport: a system where AI isn’t a separate chatbot, but a layer integrated into the operational process, available at the right time and in the right context.
The Bottom Line
88% of companies use AI, but only 6% are extracting transformative value from it. The gap isn’t technological. It’s a matter of approach: the companies winning are the ones redesigning their processes, scaling quickly, and using AI not just to cut costs, but to innovate.
2026 is the year to decide which side you’re on.
Want to understand how AI can concretely improve your technical support and maintenance processes? Contact us for a direct conversation.
