AI was supposed to make work easier. Instead, British workers are spending substantial time each week playing tech support for their digital assistants—a phenomenon researchers call “botsitting.” You know the drill: feeding context into ChatGPT for the third time, wrestling with a copilot that missed the point, or fact-checking outputs that look plausible but smell wrong.
Recent workplace research reveals the harsh math behind AI adoption. While AI tools have become ubiquitous in UK workplaces, workers are burning hours weekly on maintenance tasks that didn’t exist two years ago. The hidden cost emerges in the gap between getting AI output and making it actually usable for real work.
When Robots Need Babysitters
Workers spend days reloading context, catching hallucinations, and restarting failed AI sessions that sound authoritative but deliver unusable results.
Botsitting isn’t just occasional debugging—it’s become its own job category. Workers spend their days:
- Loading context into different tools
- Catching hallucinations that sound authoritative
- Restarting AI sessions that fail to deliver usable results
Like being stuck in Netflix buffering hell, but with quarterly reports instead of reality TV.
The grunt work is relentless: telling systems which documents matter, explaining what context they’re missing, then policing outputs for subtle errors that could derail projects. Workers have essentially become the integration layer between their organization’s scattered AI tools, filling gaps that APIs and fancy protocols can’t bridge.
The Good Enough Trap
Fatigue from constant AI supervision leads many workers to accept subpar results rather than fight for better ones.
Dr Rebecca Hinds from the Work AI Institute captures the central problem: “Adoption alone doesn’t equal transformation. If employees are spending the productivity dividend on botsitting, companies haven’t eliminated work—they’ve created a new layer of overhead.”
The fatigue is real. Rather than spend more time refining outputs, many UK workers admit they’ll accept the first AI result that looks “good enough.” That’s a recipe for gradual quality decline in everything from client proposals to performance reviews, where AI is increasingly making decisions about promotions and compensation.
Beyond the Hype
Despite claims of massive productivity gains, workplace AI often creates more overhead than genuine efficiency improvements.
The UK has embraced workplace AI adoption more aggressively than many other markets, yet the disconnect between promised gains and actual experience remains stark. The gap isn’t surprising when you realize that companies are layering AI onto existing processes without redesigning workflows or training workers properly.
This mirrors earlier digitization waves that promised efficiency but delivered email overwhelm and administrative bloat. Until organizations treat botsitting as a design flaw rather than an inevitable cost, Britain’s AI advantage risks becoming just another way to make work harder, not smarter.
