A recent Atlassian Teamwork Lab study shows that ambiguous standards—not AI itself—drive the creation of polished yet unhelpful work, or 'workslop.' Establishing clear benchmarks for quality enables teams to leverage AI tools more effectively, cutting downstream rework and restoring trust.
- Workslop occurs when AI outputs look polished but add little real value.
- Clear team standards on quality reduce workslop and build accountability.
- Atlassian pilot saw near-universal agreement on quality post-workshop.
What happened
Atlassian's Teamwork Lab conducted a pilot study involving 31 employees from its Workplace team to tackle the problem of AI-generated 'workslop'—content that appears polished but is ineffective in advancing work. The team held 60-minute workshops to help members co-create clear standards for what a high-quality page in their internal Confluence workspace should look like when produced with AI assistance.
Participants completed surveys before and after the workshops, reporting a substantial increase in clarity and confidence regarding their team’s quality expectations and AI use. Agreement that team pages met a higher-quality standard rose from 40% before the workshop to 67% after. Nearly all participants felt better equipped to give feedback and hold each other accountable to these new standards.
Why it matters
Workslop undermines productivity and trust by shifting the burden of fixing incomplete or context-lacking AI outputs onto recipients, who must interpret, check, and rewrite the content. In a prior study of 903 knowledge workers, reviewers missed critical flaws 22% more often in polished drafts, and 42% of people receiving workslop rated their colleagues as less trustworthy.
By addressing the root cause—lack of alignment on quality standards rather than flaws with AI itself—teams can harness AI’s speed without sacrificing substance. This approach moves beyond blaming AI or individual performance and focuses on shared norms to elevate collective work quality and efficiency.
What to watch next
Further research with larger sample sizes will be necessary to validate these promising pilot results and explore how diverse teams across industries can implement similar standards workshops to optimize AI collaboration. Tracking long-term impacts on productivity and trust will be key to measuring success.
Organizations adopting AI tools should consider investing in conversations and training that clarify expectations around AI-generated work. Emphasizing concrete feedback mechanisms and shared accountability can turn AI from a source of busywork into a true productivity multiplier.