A new global study from work management platform Wrike exposes why 80% of employees using AI are getting poor results. The research highlights what organisations need to do differently to unlock real competitive advantage through connected intelligence.
The momentum around artificial intelligence adoption in the workplace is undeniable. But here’s the uncomfortable truth that most leaders won’t say out loud; having more people using more AI tools isn’t making your organisation more productive. In fact, it’s making it more chaotic.
Wrike’s new research into enterprise AI adoption tells a sobering story. Over 80% of employees are now using AI at work. Yet only 34% say their teams use it in a consistent, aligned manner. The gap between adoption rates and actual effectiveness isn’t a minor problem. It’s costing organisations significant competitive advantage and creating security risks that most leadership teams haven’t fully acknowledged.
The fragmentation problem Is worse than you think
Here’s what Wrike found when surveying 1,000 knowledge workers across five countries: the tools are proliferating faster than the infrastructure to manage them.
Among employees using AI, the breakdown is clear: 53% use one or two AI tools weekly, 38% use three to five, and nine per cent use more than five. This tool overload creates predictable operational problems: learning fatigue, constant context-switching, and workflows that are actually slower than before the AI adoption began.
The root causes reveal structural failures in how organisations approach technology adoption:
- 37% have no central AI strategy; teams are experimenting independently without any unified direction
- 39% report insufficient leadership guidance; employees are left to figure out AI integration on their own
- Only a third have received company-wide training; people are deploying powerful tools without proper enablement
When you combine these factors, you get the exact opposite of what technology should deliver: siloed data, disconnected systems, inefficient workflows, and employees learning through trial and error. You’re paying for productivity tools that are generating overhead instead.
Shadow AI: The unspoken security crisis
One finding from Wrike’s research stands out as particularly troubling, the rise of “shadow AI,” a direct result of organisations failing to provide proper support and structure.
Wrike found that 42% of workers are using unapproved tools; personal ChatGPT accounts, consumer-grade software, freelance AI services. They’re reaching for these solutions to solve work problems. This isn’t employees being reckless. It’s employees being pragmatic when official channels don’t deliver.
The enablement picture explains why:
- Only 27% of organisations describe their AI efforts as “running smoothly”
- Fewer than half have company-wide training, clear policies, or structured rollout processes
- 29% deploy new tools without sufficient onboarding
When sensitive company data enters unapproved applications, organisations face real risk: data leaks, regulatory non-compliance, and cyber-attack exposure. This isn’t a technology problem. It’s a governance and strategy problem.
What employees actually want from AI (and it’s different than you think)
Wrike’s research asked knowledge workers what matters most in their AI tools. The top priorities were what you’d expect:
- Accuracy: 52%
- Speed: 47%
- Ease of use: 46%
- Privacy and security: 42%
But here’s what’s changing; employees increasingly value context awareness (26%) and role-specific relevance (28%). This reveals an important insight. Employees recognise that generic AI isn’t sufficient. Tools need to understand their specific workflows, business context, and role-specific challenges to deliver genuine value.
This connects to another critical finding: 96% of respondents believe better-integrated AI tools would be valuable, with over half stating that improved integration would genuinely transform how they work. Yet only 33% currently have strong AI integration with their core business systems. That gap represents untapped competitive advantage.
The solution: Connected intelligence, not scattered tools
Wrike’s research points to a clear answer; organisations need to stop thinking about “adopting AI” and start thinking about “integrating AI infrastructure.”
This means moving beyond scattered experimentation towards what Wrike calls “connected intelligence”; unified systems where people, tools, data, and workflows operate as one coherent infrastructure.
This looks different depending on your organisational context, but three core principles apply:
Structured data access. Siloed data is wasted data. When your AI systems can access your organisation’s proprietary data, workflows accelerate and insights deepen. This requires intentional governance, but it doesn’t have to be complex. It just has to be deliberate.
Integrated workflows. Rather than bolting standalone AI tools onto existing systems, embed AI capabilities directly into your core work management infrastructure. Wrike’s research revealed clear preference among respondents for organisation-wide AI solutions over fragmented, team-specific tools. This integration removes friction from decision-making and eliminates context-switching waste.
Proper enablement and governance. Connected intelligence means employees have clear guidance, consistent training, and secure frameworks for responsible AI use. This removes the need for shadow AI and replaces ad-hoc adoption with intentional, secure deployment.
Real-world impact: Why integration actually matters
The companies successfully executing this strategy are seeing substantial results:
- One publishing company cut proofing turnaround times by 42%
- A digital marketing agency saved 3 to 5 hours per week per employee
- A legal services firm halved the time spent in status meetings
These aren’t incremental improvements. They’re the kind of competitive advantage that shifts market position.
The Strategic imperative
The momentum around AI adoption will continue. What separates organisations that will thrive from those that will struggle is whether they can move from scattered experimentation to connected infrastructure.
For most organisations, this requires a fundamental shift; stop asking “How do we adopt more AI?” and start asking “How do we unify and integrate the AI we’re already using?”
The first group will continue drowning in tool overload and shadow AI risk. The second group will compete effectively.
Moving beyond AI pilots
Organisations like Varsity Yearbook, Jellyfish, and Kalexius have already proven what connected intelligence delivers: Varsity Yearbook cut proofing times by 42%, Jellyfish saved three to five hours per week per employee, and Kalexius halved status meeting time.
If you’re ready to move beyond scattered AI adoption and build the infrastructure your teams actually need, Wrike can help. Speak with their team to explore how connected intelligence could transform your operation.
Research Methodology: Findings drawn from Wrike’s September 2025 global survey of 1,000 full-time knowledge workers across ten industries in the US, UK, Germany, France, and Japan. All respondents were 18 years or older and employed at organisations with at least 1,000 employees globally.
Author: Thomas Scott
Bio: Thomas Scott is the CEO of Wrike. He has been with the company since 2022, having previously served as CFO.
He has more than 20 years of experience as a top executive at several companies, including Zebra Technologies, Fetch Robotics, Corning Optical Communications, and Spidercloud.
For more stories like this visit our Industrial Data & AI section
