CIOs are feeling the pressure of the AI leadership gap

As AI in the enterprise moves from pilot projects to production, many of these projects are ending in disappointment or outright failure. 

In many cases, the responsibility for these failures can be laid at the feet of CIOs, who are ill-prepared to lead organizations through the massive changes that AI brings. While AI implementations rose by 282% over the previous year, the vast majority of CIOs believe they need to enhance their skill sets to meet these AI demands, according to a 2025 survey from Salesforce. 

This is characterized as an executive "AI leadership gap," according to Wendy Lynch, PhD, founder of Analytic Translator, a corporate executive training firm.  

Lynch has a professional background in analytics and data science and has parlayed this experience into a role that she deems "an AI translator," who can help C-suite executives improve communication skills. 

"I can speak nerd and speak business," Lynch said. "Too many people are struggling to understand the implications [of the AI transition] and they need somebody who has a level of empathy as well as knowledge."  

In this Q&A, Lynch describes the nature of the AI leadership gap and what CIOs should do to better manage the massive changes that come with the AI transition. 

Editor's note: This Q&A has been edited for clarity and conciseness.

What's the role of the CIO now that AI is so pervasive in the enterprise? 

Wendy Lynch: There is a huge leadership gap when it comes to AI. All of those headlines are highlighting things that are real. I've seen roughly half -- between 40% and 50% -- of CIOs are considered AI savvy by their own CEO. In general, it depends on which survey you look at, but I would say a majority admit that they're not quite sure what to do and where to start.  

Wendy Lynch

What are some risks of having CIOs that are not AI-savvy yet? 

Lynch: It differs a lot by industry, but groups that have data that they need to be really clean can't mess up numbers. The ones that have data well-structured are in the lead, and the groups that haven't paid as much attention to the technology and the infrastructure are the ones that are lagging. A lot of the pressure right now is from boards of directors. [A recent study indicates] that only 2% of board directors as of 2025 had AI expertise. They are responding to the hype rather than the reality and putting a lot of pressure [to adopt AI], because they hear from their peers that you can save a whole bunch of money by reducing head count, so why aren't you doing that?  

Are top executives worried that the AI train is leaving the station and the company has to get on board or it will miss out? 

Lynch: They are in a double-sided challenge, because if they don't get going soon, they really are missing out on opportunities for efficiency. But if they go too quickly and they don't have the data infrastructure or the governance in place, they can cause a lot of risk for the organization. It's a tough position for them right now, if they haven't already gotten started.  

Is a lot of this being heaped on the CIO to make sense of all this and making sure you're on the right train that's going in the direction that you want? 

Lynch: There are a couple ways to think about it. Number one, CIOs most often came from a technological field. They did data infrastructure, data science or some other aspect of data work. Then they familiarize themselves with business issues, maybe got an MBA and have been leading the implementation of technology. But it's not necessarily ideal for the CIO to be in charge of this. AI is more of a strategic and sociological evolution for a company. To ask a technical person to be in charge of a big culture shift is inappropriate, but they certainly need to be on the team.  

Who else should be brought into these teams? 

Lynch: First, they have to be folks who are willing and on board. But the team should probably have some aspect of HR and some aspect of internal marketing and communication. It should have an operations leader who knows how they're doing things now and how AI changes the way that they will do things. You might want a customer-facing person on there, because it may impact what that customer gets or how they get it. The team should be clear that this is not just a side project. They have to focus on this to move the whole company forward, rather than just changing something for the sake of change. 

How can organizations start to close the AI leadership gaps? Should they bring in a chief AI officer rather than leaving it to the CIO? 

Lynch: There are a few chief AI officers being hired now, but it's not a common role yet. It's difficult for a full-time C-level professional to become an AI expert in a short period of time, but they should get some education along the way. You can have the team include a trusted individual like an AI translator or AI expert who communicates very well so that they can tell [the C-suite] what's fluff, what's real and what's scary. They should start with a business problem and have an AI expert advise them on that business problem, rather than just thinking, "We have to use this AI thing."  

Should the CIO think of implementing AI in a much more comprehensive way, rather than just a technology infrastructure project? 

Lynch: It's a mistake to have anybody -- CIO or not -- believe that implementation of AI projects is like the implementation of any other technology. This is because it has so many tentacles of what people's roles will be, where decision-making lies, how to put boundaries and safeguards around what they're doing. They want to make sure it doesn't explode because employees don't trust it or don't believe that it's in their best interest to use it. It's a process that's bigger than just switching from being a Microsoft shop to an Oracle shop. It's not just learning how to code in a different language, because it's totally changing what somebody's role might be.  

Is the problem that AI is coming in first as a technology, rather than saying there's this business problem and how can AI solve it? 

Lynch: The successful ones that I've seen have a very specific business need that causes a lot of problems and takes a lot of human energy. There needs to be some specific use cases as a starting point that are visible and trackable. It's not just, "We're going to apply AI everything, AI is going to answer all the phones or AI is going to make coding 10 times faster." Some places have decided to use AI to build all the software, only to find out that nobody knows how to fix it when things go wrong. It does make you way faster, but you could be somewhat faster in a way that all the humans are on board and know what they're doing and where the break points are, rather than writing a whole program in two days that used to take us four months.  

What do CIOs have to do now to start preparing for what's coming? 

Lynch: First, think about the business rather than about AI. They should focus on pain points within the business that are tangible and they can change using support from AI. Ask their teams for things like processes that are slow or very labor-intensive. The other thing is, the way to get educated is difficult, because there are so many things out there telling you things -- like you need to learn vibe coding or prompting. Start with case studies about how companies have successfully used AI in ways that saved money, improved safety, improved quality so they see what's possible. [You won't be successful] if you just focus on how to vibe code or make a website by just talking to AI, because there are so many things now and every day they continue to evolve. But if CIOs become familiar with how AI is being used successfully, they can start to focus on which of those capabilities are the most important for themselves in their team.  

What are the biggest risks that organizations will run into and how can CIOs help from steering the AI car into the ditch? 

Lynch: Well, there will be a lot of tow trucks pulling those cars out of ditches. There will also be quite a few companies that dabble and end up in pilot purgatory. They keep trying things, but don't really feel like it's totally ready. The next two years will be a wild ride and there will be some exuberance from people who want to do cost cutting but will cut too fast. A lot of companies may decide that they don't need any more programmers because they have AI and they'll find out that there are places where people are essential.  

Every organization needs to move pretty quickly, but I hope that they involve people who understand the culture, sociologic and people implications of it all, so that we don't harm a lot of people and businesses along the way. The CIO is going to be central to that process. They have to resist the temptation to go so far in order to meet the urgings of the board and the CEO and take some definitive steps in very knowledgeable and business-oriented ways.  

Jim O'Donnell is a news director for TechTarget, where he covers IT strategy and enterprise ESG.