Organizations are taking strategic approaches and making progress with their artificial intelligence initiatives, new research from CDW reveals. But there is significant room for improvement as they plan and prioritize their AI projects.

In a survey of IT leaders and decision-makers conducted in early 2025, CDW found 62% of respondents agreed that their organization has good ideas for AI, but they have trouble executing those ideas.

CDW experts warn that getting AI right is an essential objective for organizations across industries, as its impact extends far beyond the realm of IT. 

“This is not just a business problem. This is not just a use case problem. This is a kitchen table conversation. This is impacting our kids, our families, how we move through society, our economy, currency, trade. Everything is being impacted by this technology,” says Joe Markwith, CDW’s chief strategist for Mastering Operational AI Transformation. “It's a massive transformation beyond just the use cases and the roles that we have in our day jobs. We try to create that sense of urgency that's not just about our jobs or our businesses. It's about where we're going as people — in society and globally.”

The research suggests organizations understand the need to get started with AI. More than 98% of respondents said their organization had initiated at least one AI project, with 48% saying they had started three to five projects.

“There has been a large percentage of organizations that have been waiting to see how the market plays out, but we’re starting to see a really accelerated adoption of AI in customer experience,” says Ken Drazin, director of digital experience at CDW. “We're going to see a ton more adoption over the next couple of years, and we're starting to see that right now.”

Total Respondents: 952

More than half of respondents (53%) said they are “very confident” about their organization’s ability to plan for and implement new AI solutions. This comfort level is critical as organizations ramp up their efforts in AI. Those that don’t get started risk finding themselves far behind competitors that are learning how to take advantage of this technology.

“If you want to thrive and you want to innovate at the speed of technology, you have to consider AI, because your competitors are as well,” says Aaron McCray, field CISO for CDW. “If you don't, you may be in the unenviable position of being so far behind the curve that it's too late.”

Organizations face numerous challenges with their AI initiatives, according to our survey. These include finding the right talent and training IT teams with the right skills; having low-quality or insufficient data; and the constant pace of change in the AI space. For example, 26% cited challenges around talent and training as among the biggest pain points their organizations have faced when planning, implementing and completing AI projects.

They are unsure of whether to build their own AI capabilities, add AI tools to current solutions or deploy cloud services. “I'd say 9 out of 10 organizations that we talk to struggle with where to start, what solutions to use and how to govern and really have a lifecycle management approach to AI,” Drazin says.

Other pain points cited by respondents included issues with data quality and availability, as well as scaling and implementation. Further, 64% of respondents said they have seen a 50% return on investment or less from their organization’s AI initiatives, while 10% said they either don’t know what ROI they have achieved or that they have not seen a return on their investment.

Total Respondents: 952

CDW experts note that the AI landscape is changing more rapidly than other technologies. This can create challenges as organizations look to reach their AI goals.  Our research found that while a vast majority of organizations (75%) have a documented AI strategy, only 34% have achieved full deployment of their highest-priority AI projects.

Many organizations are turning to third-party partners to help them plan and manage their AI initiatives in an effort to overcome their pain points. Nearly all respondents (99%) indicated that their organizations have gotten help with AI from a third party. Further, 59% cited skills and capabilities as decision factors they used when selecting third-party vendors for their AI projects.

“It's relationships and trust,” Markwith says. “When you're looking for a partner, you want to be able to trust their competencies for sure.”

Total Respondents: 952

One area of the IT landscape that is seeing a massive impact from AI — both positive and negative — is cybersecurity. AI can handle data from a variety of sources and help IT teams put together a clear picture of their organization’s security posture as well as the threats it faces. In fact, 85% of respondents said they think AI can help them improve cybersecurity, and 73% said they are implementing or have implemented AI to improve security.

“AI helps put modern CISOs in a much better position to handle their roles and responsibilities,” McCray says. “The challenge has always been, do I have the right information, the right amount of information, at the right time to make decisions? With AI, we can make really intelligent decisions rapidly. I'm able to respond much, much faster and limit the exposure of impactful events, which don't become organizationally crippling because they took down my operations.”

Total Respondents: 952

Security appears to be an area where organizations are seeking help from partners for their AI initiatives. Of respondents who said they have gotten help from a third party with AI, the most common use case was security (48%).

“The easiest use case for many organizations today is modernizing the security operations center, or SOC,” says McCray. “Organizations are using next-generation AI technology to create virtual security analysts. They can literally look through all of these historical log files and start to see the correlation between events. It can speed up all of this analysis and actually speed up incident response because a lot of these toolsets are performing AI-driven security orchestration, automation and response in near real time.”

Respondents clearly indicated that AI could help them deliver a better experience for customers. When asked how strongly they felt AI can help them achieve specific outcomes, 49% said they felt very strongly that it can improve their customer experience.

“There are so many ways that AI is impacting customer experience,” Drazin says. “It can also make customer service agents happier.”

Allowing AI-enabled chatbots to handle simple customer interactions is a common way to improve the customer experience.  When asked how they’re using AI, 68% of respondents said they are implementing or have already implemented chatbots, while an additional 17% said they are planning to.  “My organization focuses on developing AI for natural language processing,” one respondent commented. “A major project involved building a conversational AI system for customer support.”

Total Respondents: 952

Chatbots provide a significant number of benefits that improve the customer experience in a variety of ways. They reduce the time customers must spend waiting for answers to most of their questions, and they often enable customers to solve their problems without needing help from a human agent. AI agents can also provide detailed summaries of customer interactions when human agents become involved, so customers don’t have to answer the same questions again.

Further downstream, AI can reduce the stress that human customer service agents experience by reducing workloads and enabling them to focus on important issues. It can also help identify recurring problems with products or services so that improvements can be made in manufacturing or delivery.

“We can use AI to look at why customers are calling your organization. It can help us understand what the top problems are that your customers are having, and we can use that information to innovate,” Drazin says. “For example, if we’re seeing a high number of returns on product X, we can tie that to manufacturing — or wherever the problem is occurring — and create processes to fix it.”

CDW’s research makes clear that many organizations are looking to get started with AI by using tools offered by trusted providers. Nearly two-thirds (65%) of respondents said their organizations have implemented or plan to implement Microsoft Copilot, while 61% said the same of Google Gemini.

“I think the adoption of those two comes down to the fact that they are solutions from providers that organizations use for productivity tools,” Drazin says. “Organizations are generally going to want to stay in that ecosystem of either Microsoft or Google.”

Using AI from a provider whose productivity tools users are already familiar with can ease adoption. IT teams can turn on the offered AI features and let users try them out.

Total Respondents: 952

A critical question facing every organization deploying AI is what infrastructure to run these workloads on. When asked how ready their organization’s infrastructure is to handle AI workloads, 35% said “somewhat ready,” “slightly ready” or “not at all” — indicating that a significant number of organizations need to improve their level of preparation.

More than a third of respondents (35%) said their AI workloads run in a private cloud, while 26% said a hybrid cloud environment and 16% said the public cloud. A cloud environment provides a solid foundation for AI initiatives, especially for smaller projects in which organizations are testing the technology on a limited basis.

“It’s really easy to do AI in the cloud for a proof-of-concept initiative because you probably don't have a high volume of users, so it takes advantage of the elasticity that public cloud offers you,” Johnson says. “That elasticity makes it extremely cheap for me to run AI versus having to buy some hardware and put it in my data center.”

Total Respondents: 952

Ultimately, organizations need to understand that AI will continue to evolve at light speed. New opportunities will arise along with new challenges, often in ways that disrupt current norms and workflows. To be successful, organizations must improve their understanding of AI and their ability to use it effectively, while also remaining agile enough to pivot when it changes.

“The transformation we’re seeing is not a point in time. We're not going to make this transformation and then be done,” Markwith says. “It's a process change. It's a people change. It's a whole new way of looking at the world around us. The foreseeable future will continue to be a moving target, and we need to embrace that disruption.”