- Only 2% of workers say AI-generated content needs no revision
- Workers spend hours each week cleaning up “AI workflow”
- AI training should be mandatory and processes should be standardized
Despite the perceived productivity benefits, many companies are spending time and money cleaning up “AI worklop”, suggesting the technology is generating a lot of unnecessary noise, new research has claimed.
Data analysis and visualizations (55%), research and fact-finding (52%), long-term reporting (52%), and content writing and marketing (44-46%) are some of the most common areas where AI tools may not be as effective as companies once hoped.
While 92% agree that AI improves their overall productivity, only 2% say AI output needs no revision.
AI is a noisy path to productivity
Three in five (58%) spend more than three hours a week revising output, with more than a third (35%) spending more than five hours and 11% spending more than 10 hours each week cleaning up generated content.
The research from Zapier adds that despite appearing polished on the surface, AI generally lacks accuracy, context or usability.
And it’s not just perception that’s down – many have experienced rejected work (28%), security or privacy incidents (27%), customer complaints (25%) and compliance or legal issues (24%).
Zapier’s data indicates two potential solutions – first, AI models must continue to improve to improve the quality of responses. But in the meantime, workers should be up-skilled to deal with AI in its current format, not what it should be.
“The companies that see the best results are not the ones that avoid AI,” explained Senior AI Automation Engineer Emily Mabie. “They are the ones who have invested in training, context and orchestration tools that turn AI from a sloppy experiment into a controlled process.”
Almost all (94%) of educated workers say AI increases productivity, but only 69% of uneducated workers agree. As a result, only 1% of educated workers say their productivity has decreased.
Looking ahead, the report calls for AI training to be mandatory for all workers handling it, prioritizing high-risk teams and tasks first. Companies can also help employees by providing quick templates and formalizing review processes.
“The solution is not fewer tools, it’s better infrastructure,” Mabie concluded.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds. Be sure to click the Follow button!
And of course you can too follow TechRadar on TikTok for news, reviews, video unboxings, and get regular updates from us on WhatsApp also.
