Work in Progress

The manager effect: Why leadership matters in the age of AI

Dale Shearin — Tue, 20 Oct 2026 06:00:00 -0700

At Dropbox, we’ve spent a lot of time learning what people need to do their best work in a Virtual First world. As the landscape continues to evolve, one question keeps us curious: What does it take to thrive in the AI era? Here, we take a look at how managers help create the conditions for success. We explore the traits and habits of those who lead our strongest contributors, and what you can learn from them to support your own team.

Recap: What makes people thrive at work?

Last year, we looked at a group we call thriving employees—Dropboxers who show both high engagement and high performance. We wanted to understand what sets them apart and the role AI plays in their daily work.

Our research showed that thriving wasn’t simply about using AI more often. The difference showed up in how these employees worked. They used AI to elevate their impact, but they also protected time for focus and recovery, invested in relationships and growth, and leaned into human capabilities like awareness, adaptability, judgment, and connection. We call these durable skills, and they become even more valuable as technology evolves.

New research: Understanding the manager effect

This year, we looked at the people who manage these successful Dropboxers. Our latest internal research suggests they share many of the same durable skills.

Managers of thriving employees:

Were 85% more likely to be thriving themselves than other managers
Were more than twice as likely to invest AI time savings in coaching, one-on-ones, and development
Had a clearer view of team performance, with 92% of our most senior managers reporting clarity on performance against expectations, compared with 74% of other senior managers

These findings raise an important question about management in the age of AI: What can we learn from managers whose employees are thriving, and how is AI changing their role?

Leading through a changing world of work

Just as distributed work called for more intentional leadership, the rise of AI calls on managers to guide how their teams learn, experiment, and use it.

But are they prepared? Recent Gallup research found a significant gap between companies’ desire to use AI and managers’ readiness to help their teams use it. While 99% of chief human resources officers (CHROs) say AI is important to their strategy, 50% aren’t confident managers can help employees use it effectively.

In the 2026 Microsoft Work Trend Index, employees whose managers actively modeled AI use reported greater perceived value from AI and greater trust in agentic AI. Microsoft also found that they were more AI-ready when their leaders created psychological safety around experimentation.

Our research rounds out this picture by looking at managers themselves and how their experience is reflected across teams. We found differences in performance, ways of working, and the employee experience.

3 habits of managers with thriving employees

1. They model strong performance and behaviors

When we drilled into the research, these leaders were more likely to be high performers (46% vs 33%), reported higher scores for making good progress on their own objectives (95% vs 84%), and reported greater satisfaction with the balance between deep focus and collaboration (64% vs 52%).

Our 2026 Life in Virtual First survey offers a closer look at their habits. Managers of thriving employees are more likely to:

Send async follow-up notes after meetings (61% vs 43%)
Avoid responding to messages during focus time or after hours (26% vs 16%)
Share meeting agendas in advance (61% vs 53%)
Collaborate effectively across time zones (61% vs 53%)

Together, these findings suggest that role-modeling matters. Leaders who practice healthy, effective ways of working make those habits feel possible for their teams.

2. They reinvest time in their people

AI can help anyone save hours. What sets managers of thriving employees apart is how they use that time. Based on our research, they’re more than twice as likely to reinvest it in coaching, one-on-ones, and team development.

Our findings also show that engaged, high-performing employees were more likely to report meaningful career development discussions (90% vs 70%) and performance feedback that helped them improve (94% vs 75%). The productivity benefit can extend beyond the manager and compound across the team.

3. They stay close to progress without getting in the way

Among our most senior managers, 92% of those with thriving employees said they understood how their teams were performing against expectations, compared with 74% of other senior managers.

That clarity was reflected in the employee experience. Thriving team members were more likely to report well-defined expectations, useful feedback, stronger development opportunities, and a sense that their perspectives mattered. Nearly all said they understood what was expected of them (99% vs 80% of others) and felt their thoughts and ideas were valued (99% vs 81%).

They also rated their managers more highly across coaching, development, and enablement, citing valuable feedback and career conversations, clear priorities, and leaders who removed barriers.

Putting the manager effect into practice

If you lead a team and want to help them do their best work in the AI era, here are four places to start.

1. Set yourself up for success

Take time to reflect on your own ways of working. Being intentional about how you manage your time and collaborate with others puts you in a better position to guide your team. Your own habits can set the tone for how everyone works.

→ Try this with your team: Define clear norms around focus time, async follow-ups, advance agendas, after-hours replies, and working across time zones.

2. Turn new capacity into better management

When AI gives you time back, consider using it toward things that matter to employees, like communication around priorities, project feedback, career conversations, and recognition.

→ Try this with your team: Pinpoint one recurring task that AI has made faster. Use some of the time it frees up for a coaching check-in, development conversation, or moment of recognition. Booking these into your calendar can help make sure you’re reinvesting intentionally.

3. Be a connector

Strong workplace relationships help people access new perspectives, expertise, and opportunities. As a manager, you can help your employees strengthen those networks through thoughtful introductions and cross-functional opportunities.

→ Try this with your team: Build a simple networking map with your direct reports. Walk through their projects together and identify the stakeholders who are important to their success. Then ask where a connection could help and make an intro.

4. Create room for agency

Thriving employees reported greater control over how they work, from protecting focus time to finding the information they need independently. Support that autonomy by making outcomes and responsibilities clear, and spelling out where they can make decisions without you.

→ Try this with your team: Ask, “Where are you waiting for my input when you shouldn’t have to?” Identify one recurring approval, decision, or handoff you can clarify or eliminate so your reports can move forward more independently.

The future manager is an architect of thriving workplaces

More than five years into Virtual First, we’ve learned that how we work deserves as much attention as the work itself. As AI changes what people spend their time on, managers have an opportunity to be more intentional about filling the space it creates.

The choices they make now will help shape what comes next. The future of Virtual First belongs to leaders who turn new capacity into greater human potential.

2026 Dropbox Trends Report: AI impact on content at work

Dropbox Team — Wed, 14 Oct 2026 10:00:00 -0700

AI is accelerating the creation of business content and raising the stakes for how that content is found, shared, and put to work. Every proposal, analysis, and project plan draws on information a business already holds and creates new knowledge for others to use. As content volumes grow, time spent gathering context, locating the right version, and moving useful output between tools can add up across teams. Our research explores how AI is changing content at work, where that friction is building, and what it means for businesses aiming to turn faster creation into lasting productivity.

Dropbox surveyed 219 professionals primarily working in technology, AI, software, finance, and healthcare at The AI Conference 2026 in San Francisco. This report brings together 70+ statistics from our latest market research on how AI is changing content at work—from time spent gathering context to the growth in content creation and where useful AI-generated work is saved and shared.

Key report findings

15.8 hours per week averaged engaging with AI at work
36.6% of AI engagement time is spent unproductively retrieving context
4.3 additional context inputs required after the first prompt
46.4% of the new work files are developed with AI
162.6% average growth in new work files created since AI adoption
93.2% of people create more work files now with AI
43.8% of useful results created by AI never leave the prompt

AI is becoming a substantial part of the workweek, and the volume of content people create is rising with it. Gathering context is a major source of friction, with people repeatedly adding files, links, and background before a result is ready to use. Useful output then spreads across shared workspaces, local files, and chats, leaving teams with a growing body of knowledge to organize, share, and put to work.

We spend 15.8 hours per week at work actively engaging with AI

What does your average workweek look like now compared with three years ago? While many professionals previously lost many hours to routine jobs reviewing, writing, researching, or developing actual code, a lot of this work is delegated to AI these days. But with AI optimizing so many areas of work, how much time do we have to invest in actively managing our AI?

The average person spends 15.8 hours a week actively engaging with AI at work alone—and this is outside of auto-running Codex jobs that operate on autopilot. But even more telling, only 2.3% of people report zero time with AI on an average week.

Here’s what people report for AI engagement time on an average week:

2.3% spend 0 hours at work engaging AI
16.4% spend 1–5 hours at work engaging AI
24.7% spend 6–10 hours at work engaging AI
7.8% spend 11–15 hours at work engaging AI
16.4% spend 16–20 hours at work engaging AI
6.4% spend 21–25 hours at work engaging AI
5.0% spend 26–30 hours at work engaging AI
21.0% spend 30 or more hours at work engaging AI

81.3% of respondents spend at least six hours a week actively working with AI, and nearly half—48.8%—spend more than 15 hours. The largest group reports 6–10 hours, while roughly one in five reports 30 or more, showing how widely the level of engagement varies.

For this group of professionals, AI already occupies a substantial part of the workweek. With that much time invested, the effort required to find information, supply context, and refine results matters to overall productivity.

36.6% of AI time is spent retrieving context

Getting AI the information it needs takes substantial time. Respondents estimate that 36.6% of their AI engagement time goes to finding, opening, uploading, or cross-referencing documents, messages, links, and information in other work apps. Applied to a 15.8-hour AI workweek, that represents approximately 5.8 hours a week spent retrieving context—time diverted from creating, refining, and completing the work itself.

Here’s how context retrieval time cost breaks down across the board:

1.8% spend 0% of AI time retrieving context
10.5% spend 10% of AI time retrieving context
22.8% spend 20% of AI time retrieving context
21.0% spend 30% of AI time retrieving context
12.3% spend 40% of AI time retrieving context
14.2% spend 50% of AI time retrieving context
5.9% spend 60% of AI time retrieving context
6.4% spend 70% of AI time retrieving context
0.0% spend 80% of AI time retrieving context
3.7% spend 90% of AI time retrieving context
1.4% spend 100% of AI time retrieving context

Nearly one in three respondents spend at least half of their AI time retrieving context, while just 1.8% report spending none. Gathering information is a recurring part of the AI workflow, and for many, it consumes a substantial share of that time.

For businesses, this points to a practical opportunity to improve AI productivity: Make relevant content easier to find and bring into the workflow. Reducing that effort could give employees more time to develop useful results and move projects forward.

Context retrieval costs for businesses

Hours spent gathering context add up quickly across a business. Applying the survey averages to all employees in a 100-person company suggests approximately 27,691 hours a year spent retrieving context—time valued at $1.15 million in salaries alone.

Reducing that time by 50% through connected context and content across the business could recover approximately 13,846 hours annually, representing $576,952 in potential productivity savings. For a 1,000-person company, that opportunity grows to $5.77 million a year.

Making relevant content easier to find and use could free up valuable time for teams. Even a partial reduction in context retrieval creates time that businesses can reinvest in customer work, project delivery, and new initiatives.

We average 4.3 extra context rounds after the first prompt

What does this context retrieval look like in practice? It continues well beyond the first prompt. People manually add new files, reference links, examples, and background information to help AI produce the result they need. Each round can require stepping away from the conversation, searching across other tools, and bringing that information back into the task.

Respondents report an average of 4.3 additional context rounds before a result is ready to use. That repeated effort adds to the time spent gathering information and puts more responsibility on the user to identify what AI needs.

The implications also extend to quality. When AI lacks relevant team and business information, users must recognize those gaps and fill them manually. Missing a key decision, requirement, or source document can leave the result incomplete or inaccurate, creating a risk of further revisions and rework.

Here’s how often respondents add context during a typical AI task:

4.1% need no additional AI context inputs
12.3% need 1 additional AI context inputs
21.0% need 2 additional AI context inputs
14.2% need 3 additional AI context inputs
11.0% need 4 additional AI context inputs
15.1% need 5 additional AI context inputs
4.6% need 6 additional AI context inputs
2.7% need 7 additional AI context inputs
1.8% need 8 additional AI context inputs
3.2% need 9 additional AI context inputs
3.2% need 10 additional AI context inputs
0.9% need 11 additional AI context inputs
2.3% need 12 additional AI context inputs
0.9% need 13 additional AI context inputs
0.5% need 14 additional AI context inputs
2.3% need 15 or more additional AI context inputs

37.4% of respondents add context five or more times after the first prompt, and 10.0% add it at least 10 times before the result is ready to use. For these professionals, adding context happens again and again throughout a task.

For businesses, this helps explain how context retrieval can consume so much AI time. The effort builds up across repeated searches and additions. Making relevant business information easier to bring into the workflow could reduce that burden and help employees reach a usable result with fewer interruptions.

46.4% of new work files are made with AI

AI is helping create the files that move everyday business forward. Sales teams use it to draft customer proposals, analysts to turn research into reports or presentations, and developers to write and revise code that used to take weeks to produce. But beyond new asset generation, AI also helps reshape existing materials—expanding a draft, incorporating new information, or adapting content for a different audience.

That contribution now represents a substantial share of weekly output. On average, respondents estimate that 46.4% of the new work files they create or save are created or substantially changed with AI. Nearly half of their new files carry AI’s influence into the work their teams review, share, and build on.

Here’s the share of new work files respondents say AI creates or substantially changes:

2.3% report 0% of new files are made with AI
9.1% report 10% of new files are made with AI
14.2% report 20% of new files are made with AI
8.2% report 30% of new files are made with AI
12.8% report 40% of new files are made with AI
18.7% report 50% of new files are made with AI
13.2% report 60% of new files are made with AI
8.7% report 70% of new files are made with AI
0.0% report 80% of new files are made with AI
5.9% report 90% of new files are made with AI
6.4% report 100% of new files are made with AI

53.0% of respondents say AI creates or substantially changes at least half their new work files, while 12.3% put that share at 90% or more. The largest single group reports a 50% share, showing how much AI impacts everyday content creation.

For businesses, this highlights the importance of managing AI-assisted work beyond the conversation where it begins. As this content becomes part of daily operations, teams need to review it, retain relevant context, and make it accessible for colleagues to use and build on.

New work-file volume is up 162.6% on average

AI can expand how much material teams produce from a single task. Research can become a report, a presentation, and a briefing document. A campaign can generate multiple drafts and versions for different audiences. A project can produce plans, updates, and supporting documentation. As that work is created and saved, the volume of files businesses need to manage can grow quickly.

Compared with before they regularly used AI, respondents report creating or saving an average of 162.6% more new work files than before they regularly used AI. Overall, 93.2% report an increase, while 3.2% report no change and 3.7% report a decline.

Here’s how respondents describe the change in their new work-file volume:

5.0% increased files 1,000% or more with AI
1.4% increased files 500% with AI
1.8% increased files 400% with AI
6.4% increased files 300% with AI
8.7% increased files 200% with AI
26.9% increased files 100% with AI
2.7% increased files 90% with AI
5.9% increased files 80% with AI
6.8% increased files 70% with AI
3.2% increased files 60% with AI
11.9% increased files 50% with AI
3.2% increased files 40% with AI
5.9% increased files 30% with AI
1.8% increased files 20% with AI
1.4% increased files 10% with AI
3.2% saw no change in files with AI
0.5% decreased files 10% with AI
1.4% decreased files 20% with AI
0.0% decreased files 30% with AI
0.9% decreased files 40% with AI
0.5% decreased files 50% with AI
0.5% decreased files 60% with AI

50.2% of respondents report that their new work-file output has at least doubled, including 23.3% whose output has at least tripled. The most commonly reported growth rate across 26.9% of people was a 100% increase in new work-file output—over two times higher than the next highest growth increase. But what really stands out is that more people report +1,000% growth in file creation with AI vs. those that reported a decline.

This rising output brings more content to review, organize, share, and maintain across the business. Multiple drafts, supporting files, and audience-specific versions can make it harder to identify the latest information and preserve the context behind it. As file volumes grow, keeping that content easy to find and ready to reuse becomes part of protecting the productivity AI can create.

43.8% of useful AI output never leaves the prompt

Once AI creates something useful, where does it go? All the powerful output is only valuable if we and others can find, review, and build on it. A proposal may need to reach a shared folder, an analysis may become part of an existing report, and a project update may enter another business workflow. But if useful results remain inside the AI conversation, colleagues will never have visibility into this work.

On average, people follow 2.2 routes for useful AI output final destinations, and unfortunately for 43.8% of us, this useful output never leaves their prompt. Shared team workspaces lead as the #1 useful AI output destination at 45.2%, closely followed by personal or local files at 44.3%.

Here’s where respondents say useful AI output ends up:

45.2% save it in a shared team workspace (e.g., Dropbox, OneDrive)
44.3% save it as a new personal or local file
43.8% leave it in the AI conversation
28.8% copy it into an existing file
25.1% enter it into another business application or workflow
22.8% share it through email or messaging
6.8% aren’t sure where it ends up

The three leading destinations are nearly tied, showing how useful AI work spreads across shared, individual, and conversational environments. That creates a potential gap between producing something valuable and making it available for colleagues to use.

For businesses, the destination of AI-generated work matters alongside the volume being created. As output grows, teams need ways to make useful results discoverable, preserve relevant context, and keep versions clear. Otherwise, valuable work can become another piece of information someone must track down before their next task can move forward.

Turning AI output into lasting business value

AI is changing both how people work and how much content they produce. Our research shows a substantial investment of time in AI, repeated efforts to supply context, and rising file volumes. Useful results then spread across shared workspaces, personal files, business applications, and conversations—creating more information for teams to manage and build on.

Together, these findings point to a growing business priority: connecting the knowledge AI needs with the work it helps create. Easier access to relevant information could reduce time spent gathering context, while better organization of useful output could help that work keep its value across a team.

Five key takeaways stand out:

AI occupies 15.8 hours a week of our time at work, on average
36.6% of AI time is wasted retrieving context, with 4.3 extra context rounds per task
$577K in potential annual productivity value from halving retrieval time across 100 employees
46.4% of new files are now shaped by AI
43.8% of useful AI output never leaves the prompt

For business leaders, the opportunity is to make the entire workflow more productive—from finding the right source material to preserving a useful result. As AI-assisted content grows, making that work accessible, understandable, and ready to reuse becomes part of sustaining its value.

Put your Dropbox project context to work with Muse, from Meta

Dropbox Team — Tue, 29 Sep 2026 02:40:00 -0700

When you work for yourself, the work behind the work adds up fast. Finding the latest brief. Chasing missing files. Organizing project materials. Getting everything ready to share. AI agents can help carry that work forward, but they’re most useful when they have the right context: the briefs, drafts, references, and project history you already keep in Dropbox.

Today, we’re launching the Dropbox connector for Muse, giving you a new way to put the files and project history you already keep in Dropbox to work.

Connect Dropbox to Muse, and your personal agent can use your project context to find relevant files, organize project materials, collect what’s missing, prepare work to share, and carry out actions as part of a broader, multi-step plan.

The idea is simple: start with the work you already have instead of rebuilding the background from scratch.

"For a small business, the context behind the work already lives in Dropbox, from creative assets to client files and agreements. Connecting Dropbox to Muse brings that trusted context into another place customers choose to work, helping them find, create, and organize what they need without jumping between tools. The opportunity is to make Dropbox useful wherever work happens, so customers can spend less time managing it and more time growing their business." — Ashraf Alkarmi, co-CEO, Dropbox

Your browser does not support the video tag.

Give Muse the goal. Start with the work already in Dropbox.

Say you run a small business and you’re onboarding a new customer. Your onboarding checklist, welcome materials, service information, and project templates are already in Dropbox. Instead of managing every step separately, you could ask Muse: "Onboard this customer using our materials in Dropbox."

With the Dropbox connector enabled, Muse can use those materials to identify the required steps, information, and project milestones. It can create a customer folder, organize the relevant documents, set up a file request for anything that’s missing, and prepare a shared link for the customer. Muse can also draft a welcome note and kickoff checklist or coordinate next steps with other connected apps, where supported.

Dropbox gives Muse more than the document that starts the process. It provides the source materials and project history behind the work, connecting what was agreed to incoming assets, new work, and final deliverables as the project moves forward. That means Muse doesn’t have to start with an isolated prompt or ask you to reconstruct the background every time.

Muse can help plan and coordinate the steps using the context available through Dropbox. For supported workflows, Muse is designed to continue working after you close the app and return when it needs your input or approval.

Build on the work you've already created

The same applies to creative work. A long-form video in Dropbox might sit alongside brand guidelines, past scripts, campaign plans, and reference assets. Ask Muse to "turn this into a week of content," and it can work from that existing body of work instead of starting from a blank prompt.

Muse can identify key themes and help prepare a video description, newsletter copy, social captions, and short-form concepts. Through the Dropbox connector, it can organize those materials into a campaign folder alongside the original recording, keeping the source, brand context, and new work together.

For freelancers and small-business owners managing several projects at once, Muse can also help create a clearer view of what needs attention. Ask it to surface missing client inputs, upcoming deliverables, and decisions that need your attention using the project materials in Dropbox. You spend less time reconstructing where each project stands and more time acting on what needs to happen next.

How Muse accesses and actions on your Dropbox content

When you authorize the Dropbox connector, you grant Muse permission to access the files and folders available through the Dropbox account you connect. Connecting Muse does not change your existing Dropbox sharing settings. When you connect Dropbox to Muse, Meta's terms and privacy policies apply to how Muse accesses and uses your content. Dropbox keeps your source files safe in your account, while Meta governs how its AI agent interacts with them.

Muse can search and retrieve your Dropbox content in the connected account and use Dropbox tools to create and organize files and folders, create shared links, and set up file requests as part of the workflows you request. Your Dropbox content should only be used by Muse in accordance with how you instruct Muse to do so.

Actions that change your Dropbox content — such as creating files, folders, or shared links, or moving, copying, or deleting content — will prompt Muse to request additional confirmation. You can review and change these permissions in your Muse account settings.

Getting started

The Dropbox connector for Muse is available in English to users in the United States and Canada. It works with all Dropbox accounts and is available across all Muse plans.

To connect your Dropbox account, log in to Muse, open the menu in the bottom-left corner, and select Settings. Go to Connectors, search for Dropbox, and click Connect. Review the connection details and click Connect again, then sign in to Dropbox and authorize Meta Muse to access your account. Once you approve access, your Dropbox account will be connected to Meta Muse.

Learn more about using Dropbox with Muse.

Need help? Visit our Help Center for additional information.

Dropbox and SpaceXAI turn project context into durable work in the Cursor Marketplace and Grok Bot

Dropbox Team — Thu, 24 Sep 2026 09:00:00 -0700

AI models and agents are only as smart as the context you give them. The latest file rarely captures the full project history stored in Dropbox: changing briefs, shifting feedback, tradeoffs, and approved decisions. Without that history, AI-assisted work can move backward by reviving rejected messaging, missing a change order, or following a ticket without understanding the requirement behind it.

We’re collaborating with SpaceXAI to introduce two new integrations to close that gap: Dropbox in the Cursor Marketplace and the Dropbox MCP for Grok Bot.

AI-assisted work should draw on the project history people already store in Dropbox, not just the last file they worked with. As the trusted context layer for AI, Dropbox connects that history to the tools people choose.

With Dropbox in the Cursor Marketplace, technical teams can use requirements, research, and decisions from Dropbox as they build. The Dropbox MCP for Grok Bot lets people give Grok Bot a role and a goal without repeatedly finding files or re-explaining the project. The underlying source material stays organized in Dropbox and existing permissions continue to apply.

Results can return to Dropbox, where they stay connected to what came before and are ready to review, share, and use again.

Your browser does not support the video tag.

Build from the context that shaped the requirements

A codebase can show how a product works, but not always why a requirement changed or which tradeoff was approved. That knowledge often lives in product briefs, API documentation, research, meeting notes, and client requirements stored in Dropbox. Without it, technical teams can lose time reconstructing decisions or risk building from an incomplete picture.

Dropbox in the Cursor Marketplace brings the materials around the code into AI-assisted build workflows. A developer can work from requirements and technical-review notes while implementing a feature. A product manager can draw on briefs, customer research, and technical decisions while drafting a specification or preparing a prototype.

They can also organize supporting materials and save new documentation back to Dropbox. When tradeoffs are made, they can record the reasoning in a README, decision log, project note, or updated specification. The integration includes 19 tools and six workflow-based skills for work like this.

Saving that reasoning to Dropbox turns it into durable documentation that colleagues, reviewers, and future contributors can use to understand how the work took shape.

“Building great software takes more than code. Teams also need the decisions and project history behind the work,” said Toni Adams, Head of Global Partnerships at SpaceXAI. “Dropbox in the Cursor Marketplace makes it easier to bring that context into the development process, capture new decisions, and keep the work connected for everyone contributing to the project.”

Your browser does not support the video tag.

From goal to action, with context already in reach

An AI assistant needs more than a goal; it needs the background behind the assignment. Campaign planners and project managers rarely work from a single file; they rely on briefs, assets, research, meeting notes, feedback, and decisions gathered over time.

The Dropbox MCP for Grok Bot makes the materials and decisions already organized in Dropbox available as the task unfolds. That lets users move directly into the assignment instead of spending the first few prompts assembling background.

A marketing lead could ask Grok Bot to develop a plan informed by approved messaging, current assets, and earlier campaign research. A construction lead could ask it to assemble an owner briefing from specifications, meeting notes, and change documentation. In each case, Grok Bot can move the work forward using relevant project history from Dropbox, not just what’s included in a single request.

The result can then be saved back to Dropbox for review, sharing, and next steps instead of remaining isolated in a Bot response.

Keep access grounded in Dropbox permissions

Your existing Dropbox permissions apply to both integrations. They can only access content the signed-in user can already access in Dropbox, and any access policies set by their organization remain in effect. Users must also authorize the connection and approve the required scopes. Any deletion must be reviewed and explicitly confirmed by the user.

As AI moves from answering questions to taking action, these safeguards keep consequential actions visible and in the user’s hands.

Getting started

Dropbox in the Cursor Marketplace
Dropbox in the Cursor Marketplace is available to individuals and organizations on all Dropbox and Cursor plans, with no current country restrictions.

Dropbox MCP for Grok Bot
The Dropbox MCP for Grok Bot is available across eligible Cursor individual, team, and Enterprise subscriptions: Cursor Pro, Pro+, and Ultra; Cursor Teams Standard and Premium; and Enterprise.

Connect Dropbox in the Cursor Marketplace to access the Dropbox MCP across both apps.

Learn more about the Dropbox and SpaceXAI partnership.

AI is the new starting point for work, but finishing it still requires a manual "last mile"

Dropbox Team — Wed, 23 Sep 2026 06:00:00 -0700

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.

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.

What we found

AI is becoming where work starts, but rarely where it finishes. More than half of respondents (54%) say an AI assistant is now their primary starting point for new knowledge work, ahead of documents, email, search engines, and project management tools. Yet only 11% say AI already generates finished work without further barriers.
The last mile remains manual. Nearly three-quarters (74%) manually move AI-generated work between applications at least three times on an average AI workday. 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. Forty-three percent say most of their weekly AI-assisted work becomes permanent team documentation or part of a shared knowledge base.

AI changed where work starts

AI is rapidly lowering the barrier to creating a first output. But starting work in AI doesn’t mean finishing it there.

Accuracy is the biggest obstacle to turning AI-generated output into finished business work, cited by 40% of respondents, followed by security requirements or company protocols at 26%. As AI-generated work becomes part of everyday business workflows, that makes trusted context, permissions, and governance an important part of getting work across the finish line.

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.

The new "last mile" of work

That work after generation is AI’s new “last mile.” Output still has to move through the context, verification, people, decisions, permissions, and workflows required to make it trusted and usable. As AI tools expand beyond generation, the challenge is keeping that work connected across the tools, systems, and people teams rely on. That’s where Dropbox comes in, connecting people and content so work can keep moving.

Consider a customer proposal. AI might generate the first draft, but someone still needs to check it against customer information, gather feedback, incorporate revisions, get approval, and save the finished version somewhere the team can find and reuse.

The next phase of AI is turning output into durable work

At Dropbox, we believe solving AI’s last-mile challenge means looking beyond generation. As work moves across AI tools, people, and systems, the content and context behind it need to stay connected, and the resulting work needs to remain trusted, accessible, and ready for teams to continue building on it.

We’re already seeing customers bring Dropbox into their AI workflows, with usage across partner AI tools growing 13x over the last four months. Through integrations with tools like ChatGPT, Claude, Gemini, Runway, Superhuman, and Monday.com, we’re making it easier to start work in AI tools while keeping it durable.

We’re building Dropbox to play a broader role in that last mile by helping content and context stay connected across AI experiences, while giving teams a consistent place to continue, share, and build on their work.

AI has already changed the first mile of work. The next opportunity is making the last mile as seamless as the first.

~ ~ ~

About the research

Dropbox sponsored a survey of 504 AI-using US B2B professionals from July 28 to July 30, 2026. Respondents represented functions including IT, engineering, sales, marketing, finance, operations, HR, and customer support.

Put your Dropbox context to work in Google Gemini

Dropbox Team — Thu, 17 Sep 2026 09:00:00 -0700

AI is only as smart as the context you give it. Too often, you’re stuck copy-pasting brand guidelines, searching through old folders, or settling for generic drafts that ignore your past work.

The result is work that's scattered and disconnected from the source material and project history you rely on every day.

That’s why we’re launching new Dropbox integrations for Google Gemini. You can connect Dropbox to Gemini to bring the context you already keep in Dropbox into your work, so you can find, summarize, and create from your own projects and past work without leaving the conversation.

We're bringing Dropbox into Gemini through two experiences: Gemini chat and Gemini Spark.

The Dropbox app for Gemini is available to anyone with a personal Google Account who is 18 or older in the US, making it easy to reference your Dropbox content while working in Gemini. For Google AI Pro and Ultra subscribers, Gemini Spark takes it further, moving projects forward as part of automated, multi-step workflows.

Together, these experiences let you bring your Dropbox files, source material, and project history into Gemini and turn that context into your next piece of work, while keeping Dropbox as the trusted home for everything behind it.

Create from your own content with Gemini

Gemini helps you use your Dropbox context to research, plan, and draft, so you can spend less time hunting through folders and more time creating.

If you’re a content creator, for example, your Dropbox already holds the working history behind your creative projects: scripts, brand guidelines, sponsor briefs, past videos, research, and photos. Instead of digging for the right file, you can ask Gemini chat to pull it up (e.g., “find the latest version of my channel style guide”), or to summarize a long sponsor brief before you start filming. Because Gemini is working from your actual files, your output is instantly grounded in your personal context, not a blank slate.

You can also turn what you’ve already made into what’s next. Ask Gemini to analyze past video scripts stored in Dropbox and pitch new angles, or draft social captions that reflect your brand voice. Rather than starting from scratch, you start from your own catalog of work, and when a draft is ready to share, you can export it to a Google Doc in a click.

Your browser does not support the video tag.

Automate multi-step workflows with Gemini Spark

Some work takes more than finding information. It takes organizing all the material behind a project, prepping it to share, and keeping the project moving. Gemini Spark helps you go beyond the conversation, working with Dropbox to organize files, generate shareable links, create file requests, and keep your work on track.

Imagine you’re a freelance designer kicking off a new project. You’ve got a brief to review, reference images and past work to pull together, and new assets still coming in. Rather than sorting it all out by hand, you can ask Gemini Spark to find and summarize the brief, create a project folder, organize your files by stage, set up a file request for missing assets, and generate a shareable link for the final deliverables.

Because Spark works with Dropbox, Dropbox stays the trusted source for your files and project work, helping keep everything from getting scattered across disconnected AI tools. Gemini’s access stays aligned with the files and folders you’re permitted to access in Dropbox. And before taking sensitive actions, like sharing, deleting, or changing permissions, Gemini asks you to confirm before it acts, so you’re always in control.

Your browser does not support the video tag.

Getting started

The Dropbox app for Gemini is available to anyone with a personal Google Account who is 18 or older in the US. Gemini Spark is available to Google AI Pro or Ultra subscribers in the US. For this initial release, the Dropbox connected app is available in English only, in the Gemini web app at gemini.google.com, and requires a personal Google Account with Keep Activity turned on.

With Dropbox connected to Google Gemini, you can put the context behind your work to use while Dropbox keeps it all together in one trusted home.

Get started by connecting Dropbox to Gemini. In Gemini, go to Settings → Personal Intelligence → Connected Apps → Dropbox to connect.

Learn more about our partnership with Google Gemini.

Need help? Visit our Help Center to learn more.

Dropbox brings trusted project context directly into the ChatGPT Library

Dropbox Team — Thu, 10 Sep 2026 12:00:00 -0700

Together with OpenAI, we’re making it easier to bring the context behind your work directly into your AI workflows.

The most useful AI work rarely starts with a single file

Imagine you’re preparing for a client review in ChatGPT. The latest brief is in Dropbox, along with the current presentation, earlier versions, client feedback, supporting research, and approved messaging. Together, those materials provide the context behind the project. But before you can get started, you’re searching for the right sources, checking versions, and bringing them into the conversation one by one.

AI becomes much more useful when it can work from the context behind a task, not just an isolated document. That’s why more customers are connecting their trusted Dropbox content to the AI tools they choose. In fact, usage across Dropbox’s AI integrations has grown more than 13X over the last four months.

We first brought Dropbox into ChatGPT through our Model Context Protocol (MCP)-powered app, making it possible to search and reference Dropbox content and take actions like saving work, organizing files, and creating links for sharing from a ChatGPT conversation.

Now, together with OpenAI, we’re making it even easier to bring the Dropbox context you need directly into your workflow.

Today, we’re introducing a new native Dropbox experience in ChatGPT

Dropbox is among the first platforms to become part of ChatGPT’s native file experience. Once connected, Dropbox is available directly within ChatGPT’s everyday file flows, giving you a faster way to find the source material behind your work and bring it into the conversation. You can browse and search Dropbox content and folders in Library, choose from recent files, @mention a file in a conversation, add it directly to a prompt, and preview files, including videos, alongside your ChatGPT conversation.

This native experience builds on the Dropbox capabilities already available in ChatGPT. Native discovery and preview make it easier to bring the right source material into your conversation, while the Dropbox app lets you take the next step by saving, organizing, and sharing what you create. The result is more than a shorter path to a file. It’s a more connected path from the context behind your work to the work you create next.

Your browser does not support the video tag.

Bring the context you need into ChatGPT

Until now, bringing Dropbox content into a ChatGPT conversation started by @mentioning Dropbox to invoke the app and asking it to find the right file. With the new native experience, you can go directly to the Dropbox file you want to use.

Once connected, Dropbox appears throughout ChatGPT’s native file experience. You can search for a specific file, choose from recent files, browse your Dropbox files and folders in Library, or mention a file directly in your prompt.

But most projects involve more than one document. A client presentation may depend on the original brief, previous versions, customer feedback, research, and approved messaging. A campaign might draw from a creative brief, concept deck, performance results, and multiple rounds of feedback.

With Dropbox in ChatGPT, you can bring relevant sources from across that project into the conversation as you work. Instead of repeatedly finding, uploading, and explaining information that already exists, you can give ChatGPT more of the context behind the task at hand.

“By bringing the files and context people already rely on in Dropbox directly into ChatGPT, we’re making it easier to move from finding information to understanding it, developing new ideas, and taking action — all in one flow,” said Sondra Batbold, Product Lead, Personalization and Context at OpenAI. “This is an important step toward making ChatGPT a more connected place to get work done, while respecting the permissions and controls people already have in place.”

Keep the source beside the conversation

Context is only useful if you can stay connected to the sources behind it. When you bring a Dropbox file into a ChatGPT conversation, you can open a preview of the original file alongside the chat. That makes it easier to ask ChatGPT to summarize a document, explain a section, compare sources, synthesize feedback, or help create something new while keeping the original material in view.

For example, an architect could open the latest project specification beside a conversation and ask ChatGPT to summarize key requirements or explain a technical section. They can refer directly to the original document as they work, verify important details, and continue without repeatedly moving between the source and the conversation.

For detailed, source-heavy work, that connection matters. It helps keep the AI workflow grounded in the material the work came from, and gives people a clear path back to the Dropbox source when they need to verify or continue the work.

Turn existing video into new campaign assets

Video can provide valuable context for creative work, and Dropbox is one of the first platforms in ChatGPT’s native file experience to support previewing video files without leaving ChatGPT. That means you can play videos directly alongside your conversation, without downloading and re-uploading them.

For example, a marketing team could preview a campaign video alongside the creative brief and approved messaging stored in Dropbox. They could then ask ChatGPT to create a new version with a branded end card and updated call to action and save it back to the campaign folder alongside the original source material.

This gives creative and marketing teams a faster way to adapt existing video for new channels, audiences, and campaigns while keeping the project context, source material, and new versions together in Dropbox.

Turn project context into durable work

Finding the right information is only the beginning. Once you’ve brought relevant Dropbox files into ChatGPT, you can use them to compare versions, synthesize feedback, identify decisions, or create something new grounded in those sources.

And the result doesn’t have to remain inside an AI conversation. When the work is ready, you can save it back to Dropbox, organize it in the right folder, and create a link to share with collaborators. Dropbox remains the home for the underlying project files and source material, while the new work has a durable place alongside the context that informed it.

For example, a creative lead could bring a campaign brief, concept deck, and client feedback into ChatGPT to prepare a review recap. Once the recap is ready, they can save it to the campaign folder in Dropbox, where the people working on the next round can find it alongside the original materials and continue from the same project context.

That creates a more continuous workflow: bring trusted project context from Dropbox into ChatGPT, use it to move the work forward, and save what comes next back to Dropbox.

AI is where work starts. Dropbox is where work becomes durable.

Keep access grounded in Dropbox permissions

Trust also means maintaining a clear relationship between AI-assisted work and the sources behind it. Existing Dropbox permissions continue to apply to this experience, so people can only discover and work with files they already have access to through their connected Dropbox account.

That means relevant project and company context can be brought into ChatGPT without changing who has access to the underlying Dropbox files. The source material remains stored in Dropbox, preserving a clear connection between the work happening in ChatGPT and the content behind it.

As AI becomes part of more business workflows, that connection becomes increasingly important: people need useful context, but they also need to know where that context came from and retain a trusted home for the underlying work.

Put your Dropbox context to work in ChatGPT

As AI becomes a place where more work begins, Dropbox is evolving from simply a place where work is stored to a customer-owned layer that helps make trusted context available to the AI tools people choose, and gives the work created with AI a durable place to live. The new native Dropbox experience in ChatGPT is another step in that direction.

You can bring relevant project context directly into your conversation, work from trusted source material, and save what comes next back to Dropbox. The underlying content remains anchored in Dropbox and its existing permissions, while new work can return to the place where it can be organized, shared, and used again.

And because people will use different AI tools for different jobs, we believe that context should remain theirs. Dropbox can provide a consistent home for the source material and resulting work even as the tools people use continue to change.

At launch, the native Dropbox experience is available on the web for ChatGPT and Work, including Enterprise audiences. Mobile availability is coming soon.

To get started, connect Dropbox to ChatGPT, then access your Dropbox files from Library.

Small businesses can move faster with Dropbox in ChatGPT

Dropbox Team — Tue, 08 Sep 2026 09:00:00 -0700

When you're running a small business, every minute counts.

Today, Dropbox is joining OpenAI's Small Business collection in the ChatGPT plugin directory as an inaugural partner. The collection features 16 essential tools chosen to help small business owners win new customers, serve existing ones, and get work done faster than ever.

Small businesses need more capacity to get the work done. In a small business, the person writing the proposal is usually the same person chasing the signed contract, fixing the invoice, and answering the customer email that came in at 9 p.m. No one hands anything off, so everything competes for the same hour.

The proposal exists in three versions in a folder someone else named. The terms are buried in an email thread. The photos the customer sent are on somebody’s phone. The job can’t start until everything is pulled back together—and next week, someone has to do it all again.

AI can help with those tasks, but only when it has the right information. For a small team, finding the latest files, gathering the relevant context, and explaining what already exists can become another time-consuming job.

With the Dropbox plugin, the files you already have become the starting point for the work. ChatGPT can find them, draft from them, save the result back to Dropbox, and share it, without anyone re-uploading.

Here's how our new plugin helps small businesses with the three jobs that come up most often: winning the next customer, serving the ones you already have, and keeping the operation moving.

Prepare the proposal without rebuilding it

A design studio bidding on a new project doesn't need to start from a blank page. The last three winning proposals, the rate card, and the scope template are already in Dropbox. In ChatGPT, ask for a draft of the new proposal using those files as the source. Review it, set the scope and pricing yourself, then save the final version back to Dropbox and send the customer a share link. No re-uploading, no copying between tabs.

Answer the customer without digging

A customer emails asking what was agreed on for phase two of a job you quoted in March. The answer spans a proposal, a change order, and a folder of notes. Ask ChatGPT to pull those files together, lay out what was agreed, and draft the reply. You can check it against what you remember, fix anything that's off, and send it.

Collect the documents you're waiting on

Onboarding a customer usually means asking for the same six things—then following up. Ask ChatGPT to create a Dropbox file request for what's missing and share the link. Everything the customer uploads lands in the right folder, ready for whoever picks up the work next.

Your permissions come with you

Dropbox sharing permissions and access controls stay the same whether work starts or ends in ChatGPT. You stay in control of decisions: ChatGPT prepares the draft or the request, and you approve what goes out.

The work outlasts the conversation

More than 700 million registered users already trust Dropbox with their most important files. Now, when you're working with AI, you can trust Dropbox to keep your content and context organized and secure—ready to find, share, and build on after the conversation ends.

Wherever you're doing your work, Dropbox remains its trusted home.

Explore Dropbox in ChatGPT. Available in the Small Business collection, on web, mobile, and the desktop app.

Updates to our Terms of Service, Privacy Policy, and Dropbox Services Agreement

Dropbox Team — Tue, 01 Sep 2026 10:00:00 -0700

We’re constantly looking for ways to improve the Dropbox experience with new feature offerings and ensure that our policies explain how our services work. So we’re letting you know about a few updates to our Dropbox Services Agreement, Terms of Service, and Privacy Policy. Here’s a summary of the key changes:

Dropbox Services Agreement

We’ve clarified your and Dropbox’s right to disclose confidential information in the event of certain types of regulatory audits.
We’ve changed the scope of your and Dropbox’s indemnification obligations in the case of misuse, and we’ve updated what happens when a customer experiences an acquisition or change of control.
We’ve clarified that Dropbox may now only be used by those over 18.
We’ve made some stylistic and structural changes to improve the clarity of the Agreement.

Terms of Service

We’ll begin allowing users to access multiple individual and team workspaces with a single account. We’ve revised our Terms to explain what it means if you use this feature.
We’ve clarified that Dropbox may now only be used by those over 18.
We’ve updated our policies on refunds, the notice period for changes to our fees, and users’ obligations if they don’t agree to updates to our Terms of Service.
We’ve changed the time period of inactivity required before we may terminate non-paying accounts to 6 consecutive months. We’ll continue to provide notice before doing so.
We’ve clarified what happens if you use your corporate email address to join a non-Team plan.

Privacy Policy

We’ve added details on how your data is handled when you access multiple individual and team workspaces with a single account, including clarifying what team admins can see.
We’ve further explained our collection and use of your data to better identify what data is being used for which purposes.
We’ve included additional details on how to appeal decisions we make about how we handle your data.

Please take a look at the full Dropbox Services Agreement, Terms of Service, and Privacy Policy, which will go into effect on January 1, 2027. You can also find our previous Terms here, our previous Privacy Policy here, and the previous Dropbox Services Agreement here. For more details, visit our Help Center.

The context layer for AI: Bringing trusted Dropbox content into OpenAI workflows

Dropbox Team — Sun, 12 Jul 2026 17:00:00 -0700

Bringing trusted content into AI workflows is already becoming part of how work gets done. As customers increasingly connect Dropbox with the AI tools they use every day, usage across our partner AI integrations has more than doubled in the past month alone, with people regularly saving AI-generated content back to Dropbox, sharing files with teammates, and organizing work from their AI conversations.

Those behaviors reinforce what we're hearing from customers. They aren't looking for another place to work. They want AI to fit into the tools and content they already trust. That's why we're continuing to invest in integrations with leading AI platforms to help customers deeply embed Dropbox into their workflows and move from isolated AI conversations to work that's easier to share, organize, and build on.

Today, we're building on that momentum with new Dropbox capabilities across OpenAI products, including new experiences in their recently announced ChatGPT Work, as well as ChatGPT and ChatGPT Codex. Dropbox customers can now use new tools to organize files and folders, create shareable links, generate file requests and execute multi-step workflows with these capabilities as official Dropbox-created skills, all directly within the OpenAI products they use most.

"AI is most useful when it has the right context," said Vibhor Chhabra, Product Lead for ChatGPT Ecosystem, OpenAI. "The Dropbox plugin helps customers combine ChatGPT's intelligence with the content they already have in Dropbox so they can spend more time on high-impact work.”

Your browser does not support the video tag.

AI workflows grounded in trusted content

As organizations bring AI into more of their everyday work, IT administrators need confidence employees are working from approved content instead of creating disconnected company information. Because ChatGPT is integrated with the Dropbox content employees already have permission to access, organizations can expand AI adoption while keeping work secure through their existing permissions, administrative controls, and governance.

Deploy AI securely: Imagine an IT administrator rolling out ChatGPT. They can give teams access to the Dropbox content they already have permission to use. Employees can find company-trusted content, generate summaries, and save AI-generated work back to Dropbox, while IT maintains permissions, governance, and administrative controls.
Launch campaigns faster: Now take a product marketing manager preparing for a major launch or customer event. Instead of searching across folders for previous launch plans, messaging documents, executive briefing materials, and campaign assets, they can quickly find the right content, generate a summary, and save that work back to Dropbox so the team can review and build on it.
Keep projects moving: The same principle applies in technical, document-heavy industries. Instead of manually gathering documentation to understand project history, a construction company preparing for a big build can work directly from the materials already stored in Dropbox. They can quickly find the latest specifications, review documentation from previous phases, understand how decisions were made, generate summaries for subcontractors, and plan next steps, all grounded in the trusted project information the broader team already relies on.

Your browser does not support the video tag.

From AI conversations to connected work

Because Dropbox is the trusted system for more than 700 million registered users and organizations, it provides the secure, portable context that powers AI across tools and models. As more work starts in AI, Dropbox helps ensure the outputs behind that work remain organized, shareable, and connected to the teams that need them long after each conversation ends.

To get started, connect the Dropbox plugin and start bringing your content into your AI workflows.