Executive summary
AI adoption has reached the point where people are getting real value from it, yet few organizations are turning those individual wins into business outcomes. Our survey of 9,500+ leaders and employees found that organizations are pushing ahead with implementation, with over a third of respondents reporting their organization has reached an operational stage where AI is embedded in select processes with defined ownership, and a quarter indicating that AI is scaled across functions or has transformed their operating models. On the ground amongst workers, while adoption is growing,1 AI maturity is a mixed picture: 64% of knowledge workers still use AI mainly for information or task assistance. The jagged frontier of AI2 has not yet leveled out; our survey found that the gap between organizational implementation and individual fluency varies widely across regions and industries.
This imbalance is amplifying fragmentation rather than resolving it. Over half of respondents struggle with information and functional silos, and over two in five cite technology silos. The consequences show up at every level. Leaders struggle with visibility into AI adoption, usage, and best practices, and a majority point to challenges with data privacy and compliance, security, and AI governance. For teams, decisions are slower and made with incomplete context. For employees, the work burden is heavier: more AI-generated output to review, more AI slop to wrangle, and more cognitive strain from context switching.
A new category of tools is emerging to close these gaps: Gartner has termed these AI workhubs, saying: “AI workhubs consolidate fragmented SaaS collaboration tools into a single, cohesive AI interface. By linking siloed data, they eliminate context switching and reduce cognitive load.”3
Two-thirds of decision-makers are interested in implementing an AI workhub within the next 12 months, and they want it to integrate enterprise data and tools, deliver enterprise-grade security and compliance, and provide granular governance and controls. Expectations are high: Leaders expect stronger team productivity and the ability to scale what works, while employees expect faster, more informed decision-making and reduced rework and context switching.
Agree AI has made individuals more productive and capable
Indicate most AI work happens in private sessions rather than shared spaces
Of leaders are interested in implementing an AI workhub in the next 12 months
AI maturity in 2026: A global snapshot
Successful AI transformation hinges on two motions — a transformation of the company’s tools and operating models, and employee adoption. Our survey took a global snapshot of AI maturity across both of these dimensions (see Figure 1): we asked respondents to self-assess their personal AI maturity, and give us their view on their organization’s level of AI implementation, using frameworks from Boston Consulting Group4 and Gartner,5 respectively. We found that:
Most knowledge workers use AI for information or task assistance. Sixty-four percent of respondents use AI for information assistance (asking questions and using the response and insights to complete tasks) or task assistance (asking AI for suggestions, checklists, and first drafts to further develop). On the other end of the spectrum, the most advanced 17% of knowledge workers have reached the stage where their workflows are partly or fully autonomous, with AI executing planned tasks with minimal user intervention.
Organizations are progressing from ad hoc experiments and pilots to operational AI. While 16% of respondents stated that their company is still in ad hoc experimentation, and 24% indicate they have moved into early pilots, over a third state their organization has reached an operational stage where AI is embedded in select processes with defined ownership. On the leading edge, a quarter indicate that their companies have scaled AI capabilities across functions or achieved full transformation of operating models and decision-making.
Where do you sit? Add your own answers
Answer the two survey questions, then compare yourself with the respondents in the charts below. Your answers stay in this browser tab and are not sent anywhere.
Figure 1: Individual maturity and organizational implementation [n=9,502]
The jagged frontier of AI: regions and industries
The jagged frontier of AI6 has not yet leveled out. Our survey surfaced significant variation in organizational and individual AI maturity across regions (see Figure 2) and industries (see Figure 3). A common thread: the regions and industries that are further ahead on organizational implementation also have the biggest gap between the average organization score and average individual score. In efforts to race ahead with company AI mandates, are these organizations focusing enough on employee enablement? A strategy leader at an aviation company captured this pitfall: “Rolling out AI tools is the easy part. The efficiency gains only show up when organizations take responsibility for bringing the hesitant majority along, and start measuring fluency instead of licenses. My own approach: launch AI lighthouse projects on real business problems, then open up the playbook in internal sharing sessions, so colleagues see it working before they are asked to change how they work.”
Figure 2: Organizational vs individual AI gap by region
Figure 3: Organizational vs individual AI gap by industry
“Rolling out AI tools is the easy part. The efficiency gains only show up when organizations take responsibility for bringing the hesitant majority along, and start measuring fluency instead of licenses. My own approach: launch AI lighthouse projects on real business problems, then open up the playbook in internal sharing sessions, so colleagues see it working before they are asked to change how they work.”
Strategy leader at an aviation company
AI has improved productivity, but fragmentation blocks organizational impact
Knowledge workers today are individually more capable than ever, as AI lets one person draft, analyze, prototype, and build at a pace that used to require a team. This is reflected in our survey, with 83% of respondents agreeing that AI has made individuals more productive and capable.
While organizations are seeing real gains in individual productivity, this has not yet resulted in better business outcomes. One reason is that adoption has scaled quickly but mostly independently, with each person having their own preferred AI tools, chat histories, and workflows: 69% of survey respondents said most AI-assisted work happens in private sessions rather than shared spaces. This explains why 73% of respondents report that speed of output has increased as a result of AI, but most did not see an improvement in decision-making or collaboration (see Figure 4). The improvement in speed is not matched with an improvement in quality, with a much lower proportion (56%) saying the latter has improved.
Indicate that most AI-assisted work happens in private sessions rather than shared spaces
“Everyone starts working towards speed as the ultimate goal. People seem to completely lose sight of what the goal is, other than making stuff, being productive. We’re putting out stuff that looks and feels convincing, but has fundamental thinking and conceptual flaws. And this is the dangerous bit — by following that route, we have collectively bypassed our own understanding of what we’re building.”
Head of Product Design at a hardware company
Figure 4: ‘How has AI changed the following aspects of work?’
At the top: misalignment and governance challenges
As AI use spreads across individuals, workflows, and departments, our research indicates that it’s amplifying fragmentation rather than resolving it. Half of respondents indicate that their organization struggles with information silos, with data being stored across disparate tools and systems, and functional silos that hinder cross-functional collaboration. Over two in five also point to technology silos, where different teams across the company have access to or prefer different tools. Leaders are well aware of this problem, with one CEO stating: “When things are scattered across Slack and docs and not well organized or something was in someone’s head and invisible to everyone else, [it is] very, very problematic.”
The proliferation of single-player AI tools and resulting fragmentation of technology, data, and collaboration have led to clear negative consequences for companies and their operating models, namely:
Misalignment leading to slow, uninformed decisions. Individuals can generate and validate ideas faster, but teams have fewer opportunities to align and build a shared view to work from. Divergence is magnified while convergence is slowed: 47% of survey respondents call out slow decision-making, and 45% point to longer review and feedback cycles. Additionally, as context and rationale stay trapped in individual chat logs, decisions are often based on incomplete or inaccurate insights. Unsurprisingly, 44% of respondents flagged that they have seen decisions being made without all the relevant context and insights.
An inability to govern and scale AI. Leaders in our survey were open about their struggle to see, govern, and scale what is working with AI to translate individual wins into organizational gains. Every organization has AI champions and power users achieving tangible gains from AI, but 40% of leaders rate their visibility into adoption, usage, and best practices as poor or non-existent. As a result, best practices and successful workflows stay hidden in individual AI chat windows and terminals instead of becoming reusable organizational capabilities. And that’s only part of the problem; between 53% and 58% of leaders highlight privacy, compliance, security, and governance challenges when implementing and scaling AI (see Figure 5).
Figure 5: ‘How challenging are the following when implementing and scaling AI at your organization?’
On the ground: increasing cognitive strain, declining well-being
For employees, day-to-day work with AI is still taking shape, and our survey points to some real friction. They report:
Higher volume of outputs and bypassed human reviews. The amount of AI-generated, often poor-quality, outputs has increased the burden of work: 59% of survey respondents indicate an increase in the volume of AI-generated outputs they need to review, and 55% note an increase in AI slop. To tackle this, some are resorting to pushing their reviews through AI. A Product Manager noted that “the outputs you would normally get from a human review have changed, and when you ask somebody to review [something], they’ll put it into their AI tool.” If critical human reviews are being bypassed due to the sheer volume of output, this further raises the risk of bad decisions.
Cognitive strain from increased context switching. The adverse cognitive impact of context switching is well documented7; our survey indicates that tool fragmentation has made this worse for many, with 44% of respondents stating that they have experienced an increase in context switching in the past 12 months. A Head of Product Design at a hardware company discussed the burden of gathering context for AI tools from disparate sources: “I’d have to export as PDFs and then group the PDFs, and then get ChatGPT to pull out the main themes of these PDFs, and it was just so laborious.”
Concerns about role blurring, job security, company culture. Anxiety about job security is a frequent, persistent headline today. Our survey unearthed some of the underlying dynamics (see Figure 6): the blurring of role boundaries and responsibilities is the top concern (60% of respondents), as is the possibility that this could lead to related roles merging (47%). The convergence of previously distinct roles could be contributing to a less collaborative company culture, a worry for 44% of respondents. On top of this, nearly one-third said social connection with colleagues is decreasing, and 1 in 5 find work lonely.
Figure 6: ‘To what extent do you agree or disagree with the following statements?’
AI workhubs are the new team homebase
Gartner predicts that through 2027, use of enterprise AI assistants and agents will create the first true challenge to mainstream productivity tools in 30 years, prompting a $58 billion market shake-up.8 Analysts highlight the emergence of a new category, AI workhubs: “commercial off-the-shelf SaaS applications that consolidate collaborative work management, workstream collaboration, visual collaboration, and common workplace productivity tools into a unified AI-driven interface.”
Our survey uncovered that decision-makers are starting to realize the potential of AI workhubs to tackle the challenges of an increasingly fragmented tool and work landscape. Two-thirds are interested in implementing an AI workhub at their organization in the next 12 months. The AI workhub capabilities they consider very important or critical include (see Figure 7):
Integration of enterprise data and tools. Underpinning the importance of relevant, timely data and context to successful team collaboration, 67% of decision-makers consider it very important or critical that AI workhubs can discover, retrieve, and surface multimodal information from any enterprise source. It’s not just data that needs to be integrated for teams to be able to work effectively; it’s also their tools: 60% view standardized protocols (e.g., MCP) for bidirectional interoperability with common work and AI tools as an important capability of AI workhubs.
Enterprise-grade security and controls. Forrester’s annual Security Survey indicates that securing AI, AI systems, and AI infrastructure is the top strategic priority for technology teams over the next 12 months.9 This is top of mind for most decision-makers in our survey, with 66% rating enterprise-grade security and compliance as very important or critical in AI workhubs. Granular AI governance and controls are also a must-have for 64% of decision-makers. A CEO of a professional services firm observed: “Companies won’t win because they have the most co-pilots. The ones that win are going to be the ones that set governance they can stand by.”
Flexibility with formats. The modern work productivity stack spans a range of formats including documents, spreadsheets, slides, and PDFs. Leaders see the capability to create editable outputs in any format as a key requirement (56%), as is the ability to do high-fidelity conversions between formats (47%).
“Companies won’t win because they have the most co-pilots. The ones that win are going to be the ones that set governance they can stand by.”
CEO of a professional services firm
Figure 7: ‘How important are the following AI workhub capabilities for your organization?’
A digital canvas is the critical foundation for AI workhubs
Gartner notes that multimodal human collaboration facilitates real-time and asynchronous collaboration within a central canvas, featuring rich presence, granular version control, and comprehensive change management for transparent, accountable co-authoring.10 Decision-makers in our survey recognize the benefits: 53% say that real-time and asynchronous collaboration on a multiplayer canvas is a very important or even critical capability in an AI workhub.
of leaders consider real-time and asynchronous collaboration on a multiplayer canvas a very important or critical AI workhub capability
“If we as a group on the canvas are looking at these outputs together, and criticizing them, and analyzing their validity, their value, how well they align with what we’re trying to do as an organization — all of these discussions are really valuable to have as a group. And it’s a nice scaffolding to do that.”
CEO of a professional services firm
Leaders and employees expect collaboration, productivity, and governance benefits from AI workhubs
Respondents see AI workhubs as a way to solve one of their main pain points in the AI era, with over half expecting faster, more informed decision-making at their company. Other expected benefits include improved collaboration and better alignment, with one Product Manager stating how working on Miro, an AI workhub, meant that, “At the end, there was no surprise for anyone when we released something; it is exactly what they saw and what they expected.” Respondents also expect reduced context switching and rework, pain points that a significant proportion of knowledge workers currently experience.
Business leaders specifically are optimistic about the benefits of AI workhubs beyond individual gains: 54% expect improved individual as well as team productivity, and 43% expect better team and cross-functional collaboration. A Client Partner highlighted the team cohesion enabled by AI workhubs: “The best parts of a tool like Miro is where you have the real-time collaboration, real-time feedback loop, and also some interactivity that’s fun for people.” Leaders also expect to be able to tackle one of their main challenges with AI adoption: the ability to scale AI workflows and best practices (50%).
Recommendations
Appendix
Appendix A: Methodology
In this study, Miro conducted an online survey of 9,500+ leaders and employees from cross-industry product development, IT, operations, and other functions based in North America, Latin America, Europe, and Asia-Pacific to understand the current state and challenges of AI transformation at their companies, and their views on AI workhubs. Respondents were offered a small incentive as a thank-you for time spent on the survey. The study began in September 2026 and was completed in October 2026.
Appendix B: Demographics
Composition of the full survey sample (all respondents).
Survey sample composition
Appendix C: Notes and sources
1. Source: “Global Workforce Hopes and Fears Survey,” PwC, 29 September 2026 ↩
2. Source: “The Shape of AI: Jaggedness, Bottlenecks and Salients,” Ethan Mollick, 20 December 2025 ↩
3. Source: “Innovation Insight: AI Workhubs,” Gartner, 11 August 2026 ↩
4. Source: “The AI Adoption Puzzle: Why Usage Is Up But Impact Is Not,” Boston Consulting Group, 8 December 2025 ↩
5. Source: “AI Maturity Assessment,” Gartner, 8 June 2026 ↩
6. Source: “The Shape of AI: Jaggedness, Bottlenecks and Salients,” Ethan Mollick, 20 December 2025 ↩
7. Source: “Multitasking: Switching Costs,” American Psychological Association, 20 March 2026 ↩
8. Source: “Innovation Insight: AI Workhubs,” Gartner, 11 August 2026 ↩
9. Base: 2,903 security decision-makers | Source: Forrester’s Security Survey, 2026 ↩
10. Source: “Innovation Insight: AI Workhubs,” Gartner, 11 August 2026 ↩
