AI Complexity Straining UK IT Teams, Freshworks Study Finds

What happened

New research from Freshworks indicates that artificial intelligence is adding to the workload of UK IT teams rather than easing it, according to reporting by CX Today. The study found that staff tasked with maintaining AI systems are spending significant time monitoring outputs and correcting errors — precisely the kind of manual effort these tools were meant to eliminate.

The findings suggest a gap between the promise of AI-driven efficiency and the operational reality inside many UK organisations, where added complexity is shifting effort rather than removing it.

Why it matters

For technology and experience leaders, the research is a reminder that deploying AI is not the same as simplifying operations. When systems require constant human oversight to catch errors or unexpected behaviour, the promised productivity gains can be offset — or even reversed — by new categories of hidden work. That has direct implications for service quality: IT teams stretched thin on maintenance have less capacity to focus on the systems and journeys that shape customer and employee experience.

The findings also point to a broader governance question for AI adoption. Organisations investing in automation need to build in monitoring, quality assurance and escalation processes from the outset, rather than treating AI as a "set and forget" layer. Without that discipline, complexity accumulates quietly until it surfaces as slower service, inconsistent outputs or frustrated staff.

The Renascence take

This is less an AI failure story than a service-design failure story. The tools are doing what they were built to do; what's missing is the operating model around them — clear ownership, defined thresholds for human intervention, and realistic expectations about where automation still needs a safety net.

Most organisations treat AI rollout as a technology project when it's really a workforce redesign project. The real cost isn't the software — it's the invisible labour of babysitting systems nobody fully trusts yet. Before adding another AI layer, leaders should map exactly where human correction is happening today, because that's where the next investment should go: not into more automation, but into the governance and training that make the automation you already have trustworthy enough to leave alone.