• 2 min read
AI fatigue is a design failure, not a worker problem
A TechRadar Pro Perspectives piece argues that rushed AI rollouts are adding admin and burnout, with only 31% of UK businesses using multi-agent workflows.
Image: TechRadar
If AI is making employees more tired instead of more productive, the problem is likely the system around it, not the workers using it. That is the core argument in this TechRadar Pro Perspectives article, which points to poor implementation as the cause of so-called “AI brain fry” — the cognitive fog and loss of concentration that comes from constantly supervising AI tools.
The piece says 69% of UK businesses are now implementing AI assistants, but most have not built the operating structure needed to make those tools genuinely useful. Instead of removing repetitive work, many deployments are creating new layers of it: writing prompts, checking whether responses are reliable, interpreting outputs, and correcting errors.
Only 31% of businesses are using multi-agent workflows, according to the article. That leaves many employees doing the orchestration work themselves rather than letting systems route tasks, validate outputs, and manage handoffs between AI agents. In the author’s framing, that is the difference between employees using AI and managing AI.
Recommended reading
AMD Helios packs 72 GPUs into one rack
What an adaptive operating model looks like
The article argues that UK businesses need to move beyond simply deploying tools and instead build an adaptive operating model. That means clearly defining:
- which tasks AI can handle autonomously
- where human judgment remains essential
- how work moves between people and systems
In practice, that could mean AI agents taking on first-pass research, data synthesis, and draft outputs, while humans focus on direction, judgment calls, and exception review rather than checking every result.
The piece also argues that businesses need stronger domain expertise and technical literacy so employees can challenge AI outputs critically and design workflows that are reliable, not just workable. If an AI system needs constant human supervision to stay dependable, the article says it has not been properly engineered.
Employment law raises the stakes
The article also ties AI design choices to legal risk. It notes that the UK government’s Employment Rights Act 2025 is set to reduce the qualifying period for unfair dismissal claims and remove the compensation cap from January 2027. That, the author argues, will increase the human and legal costs of poorly managed AI-driven workforce change.
TechRadar notes that the article was published as part of TechRadar Pro Perspectives and that the views expressed are the author’s, not necessarily those of TechRadar Pro or Future plc.
Enterprise Editor
Marcus follows the money. He covers enterprise software, cloud architecture, and the tectonic shifts in Big Tech strategy. He translates dense earnings calls and complex M&A activity into actionable insights about where the industry is actually heading. If a tech giant makes a silent pivot, Marcus is usually the first to notice.
via TechRadar
