A Dropbox-sponsored survey of 504 US professionals finds AI is now the primary starting point for many knowledge tasks, but accuracy, security, permissions, and moving content between apps keep most outputs from being finished work.
54% of surveyed professionals now begin knowledge work with an AI assistant, but only 11% say AI produces finished work without further barriers.
Accuracy (40%) and security or company protocols (26%) are the top obstacles to turning AI output into trusted, final work.
43% report that most of their weekly AI-assisted outputs become permanent team documentation or part of a shared knowledge base.
Thursday, September 24, 2026 A Dropbox-sponsored study of more than 500 US professionals who use AI reveals just how wide the gap is between where work starts and what it takes to finish it. AI can generate an answer in seconds, but getting that output across the �last mile� into finished work still requires the right context, verification, people, permissions, and workflows. Today, that often means moving content between apps, sharing it with team members, and getting it ready for review.
Compare Services FAQ All WebPost WebRelease Targeted Media Wire Service AI is the new starting point for work, but finishing it still requires a manual "last mile" AI is supposed to reduce busywork, but people are still doing much of the manual work required after AI creates an output. A Dropbox-sponsored study found that while 54% of professionals now start knowledge work with AI, 74% manually move AI output between apps at least three times on an average AI workday, and only 11% say AI produces finished work without further barriers. For 30%, that happens six or more times, while 65% use three or more applications to complete a typical AI-assisted task.
AI output is becoming organizational knowledge. Starting work in AI doesn�t mean finishing it there. That helps explain why 84% of professionals surveyed agree that AI is helping employees create work faster than their organization�s processes and workflows can keep up.
You sat down to finish one task. The chat helped. Then it offered another angle. Then a cleaner draft. Then a “we could also do this.” You kept going. Of course you did. The next version was one click away. Does this sound familiar to you? That’s the perfectionism trap with AI. The friction is […]
I keep coming back to the same thing when people ask me how to start with AI. The tools aren’t the hard part. Going from zero to one is. Once you use it for a single, low-stakes task, the rest gets easier. Until then, it sits on your phone like every other app you downloaded […]
TL;DR: The AI Workflow Inventory Finalizes the A3 Delegation System You probably know your own AI shortcuts: the colleague who drafts the stakeholder update, the nightly job somebody set up before they left, the interview notes that go through a model on demand. However, I want to challenge you: that perceived knowledg
The article discusses the integration of artificial intelligence (AI) in small businesses, particularly in construction, emphasizing the importance of identifying and measuring administrative tasks that delay project timelines. It outlines five steps for successful AI adoption, highlighting the need for clarity in proc
The early AI era is over. The established AI “brain” is about to get a “body.”
AI made producing work nearly free, so the job is absorbing complexity before handing work to someone else, with examples from tickets, PRs, and slide decks.
Loading more related stories...
Open the app view to save this story, compare related coverage, and continue from the same source.