Implementing AI at Work? Here’s How to Cut Risks and Boost Performance

Despite workers’ perceptions of the benefits of AI implementation, there remains a gap between its use and business-level measurement.

As per CDW’s 2025 Modern Workspace Report, only about one-quarter of employees believe that AI helps to reduce costs (28 percent) or leads to revenue growth (23 percent).

On the other hand, executives have a higher perceived value of AI. As per the report, executives view areas such as productivity (61 percent) and improved output quality (54 percent) as direct benefits of AI.

This disconnect shows that while employees see AI as a tool for task efficiency and productivity, many organizations may not be fully demonstrating how AI use can positively impact overall business outcomes.

This contrast in AI’s value perception and realization highlights the need for improving how AI is implemented to have a more significant impact within organizations.

As Brian Matthews, Head of Modern Workspace Services, CDW Canada, pointed out, “We are seeing increased AI adoption but the perceived value of AI is not matching the adoption.

The following risks may prevent organizations from obtaining full value from their AI initiatives, as described below.

1. Sensitive corporate data exposure

The 2025 Modern Workspace Report reveals that 49 percent of employees state sensitive data exposure through use of free unauthorized AI tools as a concern.  In the absence of data security techniques, sharing sensitive information with an AI chatbot can prove to be risky.

The shared data can be leaked on the internet, leading to theft or misuse, which creates reputational and compliance risks.

2. Lack of regulations regarding AI tools

With 52 percent of employees citing unclear regulations as a concern, organizations must prepare for a landscape where governance is still evolving. AI for work must be implemented with regulations around access control, authorized use and data safety.

“While organizations are keen on AI, they are not giving their planning enough weight. They need to understand the opportunity that's ahead of them and how to extract full value from AI,” said Matthews.

3. AI responses and model reliability

AI models deployed for office work come with the risk of hallucinations or imperfect responses. The report found that 36 percent of respondents find the reliability and accuracy of AI-generated content as a concern.

This can lead to trust gaps and misalignment with the needs of workers, who might find AI tools unfit for productive outcomes.

4. Lack of worker training and knowledge gaps

The report highlighted that a majority (67 percent) of employees using non-approved tools learn AI through trial and error. This may lead to gaps in the employee understanding of AI-borne threats while reducing the effectiveness of AI for work.

The same disparity can be seen in utilization of AI tools, as 51 percent of employees state not using AI more often due to underperformance and lack of task relevancy.

The following strategies can help IT leaders build a culture of AI-powered productivity by addressing implementation gaps and risks.

While task-level automation such as email drafts and meeting summaries can be useful for employees, they don’t reflect the full capabilities of AI.

Organizations need to dive deeper into their workflows to find use cases with a much larger AI impact.

Identifying high-impact use cases means looking at processes that truly influence business outcomes such as renewals management, customer engagement or compliance reporting.

As per the CDW Modern Workspace Report, 52 percent of respondents admit using non-approved tools at work weekly.

This can create security and safety risks for employees at the workplace. IT teams may not have full visibility into non-approved tools, making them hard to regulate. And if an employee inadvertently shares sensitive data with an AI bot, it can lead to data security violations.

Therefore, IT decision-makers must work towards setting AI usage policies that can curb unmonitored AI access. As per CDW’s Modern Workspace Report, 78 percent of employees are comfortable using AI when organizational policies are in place.

With increased employee confidence flowing from formal policies, it becomes easier to drive adoption and reduce the risks of possible security vulnerabilities.  

“With AI, it is not just about finding the right tool and putting it into an environment. It’s also spending time and effort in alleviating the concerns and barriers,” said Matthews.

Here are three policies that IT leaders can implement for safer use of AI:

• Data classification policy: Define what categories of data (e.g., confidential, customer, public) can or cannot be shared with AI tools.
• Approved tools list: Specify which AI platforms are sanctioned for workplace use and prohibit unapproved public tools.
• Access control and permissions: Apply least-privilege principles to data resources that commonly interface with AI tools.

As per the CDW Modern Workspace Report, 67 percent of employees learn AI through trial and error, which highlights a gap in formal employee trainings.

This gap can be closely tied to another finding of the report, which states that 51 percent of employees are not using AI more often due to underperformance and lack of task relevancy. 

The lack of proper employee training leads to employees having limited knowledge of AI tools, which weakens their ability to use AI constructively.

Even while using regulated AI tools, employees are often unable to maximize the benefits.

This is a key opportunity for IT leaders to transform how their employees learn to use AI and develop the skills necessary for AI-driven productivity. As Matthews puts it, “We should see formal training at the top of the list. We should establish formal training on AI and then encourage trial and error, encourage going to the platform that users learn the best from, with guardrails in place.”

When building training programs for employees, IT leaders should prioritize the following:

• Role-specific learning paths: Tailor training to how different teams (e.g., finance, marketing, operations) can use AI in their daily workflows.
• Hands-on practice with approved tools: Provide guided exercises on work-approved AI platforms so employees build confidence in real-world tasks.
• Ongoing upskilling and support: Move beyond one-time sessions by offering continuous learning resources, refreshers and access to expert support.

With training in place, employees experience higher productivity, greater trust in AI outputs and faster realization of value. As per the report, 75 percent of employees feel comfortable using work-approved AI tools with access to formal trainings.

A striking insight revealed in the Modern Workplace Report is that only 23 percent of IT decision-makers with approved AI tools have engaged third-party consultants to support implementation.

This low number is worth noting because while AI adoption is climbing, many organizations are navigating AI deployment in silos, without the expertise needed to manage governance, risk and long-term ROI.

As IT leaders plan to scale AI adoption in their organization, they are often met with challenges across technology, compliance, user expectations and workflow optimization.

These challenges could make it difficult to handle AI adoption without expert strategies and may lead to potential downsides, such as:

• Longer time-to-value with resources spent on misaligned pilots
• The risk of creating security vulnerabilities while implementing AI
• Lack of education around AI best practices, which prevents it from being used most productively

IT leaders can address this by bringing in trusted AI experts like CDW who provide structured advisory services, build tailored adoption roadmaps and support training and governance frameworks.

With this, they can accelerate safe implementation, reduce the burden on internal IT teams and help generate measurable business impact.

As IT leaders look for ways to maximize the value of their AI investments and enable more employees with useful AI tools, CDW’s AI expertise can play an important role. 

CDW Canada’s AccelerateAI Services are designed to help organizations move beyond experimentation to secure, enterprise-ready adoption.

With offerings that range from proof-of-concept pilots to full-scale deployment, CDW can help ensure your AI project is aligned with real business priorities. Our experts help IT leaders identify high-impact use cases, establish governance frameworks, measure ROI and deliver training programs that build workforce confidence.

Backed by Microsoft-certified specialists and proven frameworks, CDW enables Canadian organizations to scale AI responsibly, securely and with measurable business outcomes.