Wrike Research Finds Disconnected Workflows Limiting AI’s Potential

In a new global research report titled *The Age of Connected Intelligence*, Wrike set out to investigate how enterprises are using AI tools at work — and whether they are getting what they expect out of them. The headline finding: while AI usage among knowledge-workers is high, many organizations are still seeing limited returns because their workflows remain fragmented and disconnected. Without integration and shared context, AI tools are simply recreating the same silos and inefficiencies they were designed to eliminate.

GanttPRO is the best project management tool you can find!

High AI uptake, but limited integration

According to Wrike’s survey of 1,000 full-time knowledge workers across the globe, 82 % report using AI tools in their work.

Of those, a majority (53 %) say they use between one and two tools on a weekly basis. Yet, despite this high uptake, large gaps remain in how these AI tools are embedded into day-to-day work. A striking 96 % of respondents say that AI tools which share context across systems would be valuable; and 51 % believe that such connected tools would transform how they work.

In short: organizations have focused on deploying AI, but fewer have advanced to a point where AI is seamlessly woven into workflows, aligned with systems and data, and helping people collaborate more efficiently.

The gap between experimentation and operationalization

The report highlights what may be a phase transition in enterprise AI: plenty of experimentation, but far fewer fully integrated deployments. One consequence is the rise of “shadow AI” — when employees adopt AI tools outside formal corporate policies or governance. Wrike’s survey found that 42 % of workers have at some point used AI tools that their organization did not officially approve.

Even more telling: 20 % say their organization has not officially rolled out any AI-approved tools, and 15 % are unsure which tools are officially supported.

This “shadow AI” phenomenon signals a disconnect: the eagerness to use AI is there, but the structures — training, policies, governance, integrated tools — are not yet consistent or mature. Employees want to leverage AI — but they also want clarity, support, and guardrails so they can use it confidently and responsibly.

Structure, strategy and governance matter

Wrike’s findings suggest that many organizations are still in the early stages of their AI journey. While 46 % say their company is “making progress” with AI, only 27 % claim their efforts are “running smoothly.” Fewer than half of organizations have comprehensive training in place, role-specific enablement, or company-wide policies covering AI usage.

When asked what would enable AI to deliver more value, workers identified several key areas:

• Stronger integration between AI tools and core business systems
• Unified platforms that connect tools with workflows
• Clearer training, communication and governance around AI strategy

In other words: the value of AI doesn’t simply come from deploying tools — it comes from embedding those tools into the way work really gets done, with shared context, data flows, common platforms and trained people.

The concept of “connected intelligence”

Wrike introduces the idea of “connected intelligence” as the next wave of enterprise work: one in which people, systems and data are connected and coordinated, and AI tools orchestrate across that network. The research shows momentum: 95 % of respondents said they’d delegate at least one task to an AI agent today, and 90 % say it would be valuable if those AI agents could coordinate across multiple tools that they already use.

The implication: organizations that move from isolated AI usage toward connected, orchestrated workflows will likely differentiate themselves from those still dealing with fragmented tools, unclear governance and uncoordinated efforts. As Wrike’s CEO, Thomas Scott, put it: “The race isn’t just to adopt AI now… It’s to connect it.”

Preferences for AI deployment models

The survey also explored how organizations might approach AI deployment at scale. Among respondents:

• 34 % preferred an organization-wide AI solution, one consistent across teams.
• 25 % expressed interest in team-specific tools (i.e., tailored per department).
• The largest group, 41 %, preferred a hybrid approach that offers both organization-level scale and team-level flexibility.

This spectrum reflects the trade-offs: uniform platforms offer consistency, governance and scale; team-specific tools offer agility and customization. A hybrid model may give the best of both, provided integration and oversight are built in from the start.

Why disconnected workflows hold AI back

From the research, key reasons disconnected workflows limit AI’s potential emerge:

• Redundant effort: When AI tools are siloed from other systems and workflows, they can generate outputs that don’t easily integrate, requiring manual hand-offs or repeated work.
• Lack of context: AI models need rich context — data about task, workflow, history, role — to deliver relevant results. If that context is missing, the AI’s value drops.
• Fragmented data: Data spread across multiple platforms without integration means AI can’t easily draw from the full picture of work, decisions and outcomes.
• Shadow AI risks: When employees adopt tools outside central governance, it raises risks around data security, compliance and duplicative work.
• Change management gaps: Tool-deployment without training, process redesign or governance leaves employees unsure how to use AI effectively or safely. Thus, even though many employees are using AI tools, the full value is stalled by the lack of connective tissue across tools, people and processes.

What organizations should do next

Based on these findings, Wrike recommends that organizations seeking to unlock AI’s full value should focus on building connected intelligence to support modern work. Some key actions include:

• Adopt unified platforms that connect tasks, systems, data, people and AI tools — so workflows carry context across the lifecycle of work.
• Embed AI into work, not just bolt it on: Rather than stand-alone pilots, ensure AI is linked into the workflows people actually use, with meaningful roles and data flows.
• Invest in governance, training and enablement: Clear AI policies, training for employees on how to use AI, guidelines for approved tools, roles & responsibilities.
• Balance scale and flexibility: Use a hybrid model — organization-wide platforms provide consistency and governance; team-specific solutions provide agility.
• Measure and evolve: Monitor how AI tools are being used, the outcomes they generate, and continuously refine workflows, roles and tool integrations. Companies with mature, connected AI strategies will be better positioned to drive efficiency, coordination and innovation — rather than simply layering new tools on top of old processes.

Conclusion

In short, Wrike’s research surfaces a pivotal moment for enterprise AI: the shift from experimenting with many separate tools toward creating workflows and platforms in which AI, systems and people are deeply connected. While AI adoption is widespread among knowledge workers, the full promise remains unrealized because many organizations have yet to break down silos, integrate context and provide the structured environment needed for AI to really flow.

The message is clear: deploying AI tools is only the beginning. To reap the benefits, organizations must invest in the underlying work architecture — people, processes, data, systems — and seek to build “connected intelligence” rather than isolated digital islands.

For organizations planning their AI strategies heading into 2026 and beyond, this means shifting focus from “which tool” to “how it fits into our workflows, platforms and context.” Only then can the promise of AI move from hype to real productivity and transformation.