A snapshot of the AI in IT landscape in 2024, presented in partnership with ITSM.tools. Learn what your peers are doing, optimistic, and cautious about while adopting AI in IT.
As a follow-up to a global 2023 IT service management (ITSM)-focused survey on artificial intelligence (AI) adoption in IT organizations that showed significant growth in AI interest and use, two separate surveys focused on the state of AI adoption in North America were conducted in November 2023.
The first survey posed AI-related questions to IT professionals, and the second to end-users.
It can be hard for organizations to understand whether they’re ahead of or behind the AI adoption curve. To help, the first “core” survey question asked the respondents to describe their organizations’ current AI adoption state.
Although, the selection of some of these options might have been dependent on respondent interpretation. Eg. whether the addition and use of AI-enabled capabilities within existing SaaS software was considered AI adoption.
While IT’s adoption of AI-enabled capabilities might seeminevitable, it’s interesting to understand the origins oforganizations’ investments.
The IT team was the originator of AI adoption activities innearly two-thirds of organizations (61%), with the C-suiteaccounting for one-quarter (24%).
When the “Not applicable” and “Other” responses are removed, the “inflated” percentages are 72% and 28%, respectively – which better shows the difference between these two drivers of AI adoption in IT.
Looking beyond these full sample figures, organizations where the C-suite had originated the need for AI had progressed less than those where the IT team had done this.
The lack of appropriate AI skills and resources is often cited as a barrier to AI Adoption. A surprising 28% of respondents stated that their organizations don’t have anyone specifically focusing on AI.
However, the majority of these organizations (94%) hadn’t progressed with AI in production. At the other end of the spectrum, 65% of organizations have two or more people focused on AI, and another 7% have AI-focused resources within a separate function.
Looking beyond these full sample figures, where the AI team sits within another function, close to two-thirds (61%) of the organizations were at the Partial Integration stage, with the remainder in Pilot Projects and Early Exploration. None were at Full Integration.
The extra costs associated with AI adoption are inevitably a barrier to its adoption. However, more than half (60%) of respondents stated that their organizations are spending at least 5% of their IT budgets on AI. When the data is limited to only the organization with AI spend, this is 74%.
Looking beyond these full sample figures, 87% of the organizations with no IT budget allocation don’t have anyone specifically focused on AI. None have either a Partial Integration or a Full Integration.
As with any new technology, it’s important to understand its benefits to business operations and outcomes.
The three most stated benefits of AI adoption in IT are:
These top benefits were as expected, and only 7% of them didn’t think that AI would benefit their organizations much.
For the survey question related to the speed with which IT organizations would adopt AI, it’s important to appreciate that it could be interpreted and responded to in two distinct ways – the respondent could answer for their organization (and potentially others they know), or they could answer for the IT industry as a whole.
The highest scoring response (before rounding) was that AI will take 1-2 years to become common within IT teams (28.4%). However, 27.6% of respondents believed this was already the case – logically, this is likely to reflect the state of AI adoption in a finite number of organizations rather than the IT industry as a whole.
Only 20% of respondents thought that AI would take over two years to become mainstream in IT.
A common challenge or barrier to AI adoption has long been the availability of suitable people and expertise. This was joint fourth in this survey, behind customer data security (42%), additional cost (39%), and inaccuracy or inconsistency (33%), and tied with governance and compliance (28%). The lack of fit-for-purpose AI tools was stated by only 10% of respondents.
The customer data security challenge was most prevalent in banking and finance, software, manufacturing, and retail and e-commerce. The same was true for the additional cost challenge, too.
While AI is seen as beneficial to IT organizations, the final “core” survey question also asked about the areas that respondents would not like AI to penetrate.
The three top most stated areas were:
5% of respondents were happy for AI to penetrate all of the area options.
This is another question where the potential for response misalignment needs to be recognized. Particularly whether the respondents considered their answers in the context of AI operating with minimal human intervention or augmenting existing human activities and capabilities.
81% of the respondents in the survey sample had contacted their IT support team at least once in the past six months.
25% of the respondents had contacted IT support three or more times and only 19% of them had never contacted IT.
Roughly 8% of the respondents didn’t have an IT support team. The data points shared above take into account only the ones who did.
The most common (primary) methods of accessing support are:
Interestingly, the IT support portal only accounted for 6% of primary contact methods – perhaps a sign of question interpretation and its use for service requests but not for issue handling.
Looking beyond these full sample figures, the respondents who stated they had contacted IT support three or more times in the last six months were less likely to call (only 17% versus 25% for the full sample).
Only a quarter (26%) of survey respondents were happy with how their IT support team worked. The most common improvement requests were:
When the “happy” respondents are removed from thesample data, these percentages increase to:
Looking beyond the full sample figures, the respondents who stated they had contacted IT support three or more times in the last six months were far less likely to be happy with IT support (only 13% versus 26% for the full sample), with the largest improvement deltas for this group.
Interestingly, the largest deltas were for the three highest “full sample” improvement areas along with shortening the response and resolution times. At the other end of the delta spectrum, the high-contact users weren’t any more interested in fixing their own issues or knowing where they are in the process/queue than the full sample.
While it’s appreciated that end-users might not know whether their IT team is using AI for support or not, the results of this survey question were positive in favor of AI use. 36% were happy with their IT teams’ use of AI, and another 19% would like them to use it. When the “Don’t know” responses are removed, this is 46% and 25%, respectively, and 71% in total.
Looking beyond these full sample figures, the respondents who had contacted IT support three or more times in the last six months were far more likely to have access to helpful AI (43% versus 36% for the full sample).
While correlation does not imply causation, this could be indicative of making IT contact easier resulting in greater IT support use.
75% of survey respondents stated that they’re already using free AI tools like ChatGPT for their work. Close to half (46%) of the respondents use these tools at least once a week.
Looking beyond these full sample figures, the respondents who stated they had contacted IT support three or more times in the last six months were far more likely to use AI tools on a daily basis (28% versus 18% for the full sample).
We asked end-user respondents what they used free AI tools for.
The most common AI tool use cases for the full sample were:
Interestingly, 72% of the survey sample had heard of ChatGPT versus the next highest option, Bard, at just 4%.
39% of survey respondents weren’t concerned about their organizations’ use of AI. When the “Don’t know” and “We’re not using AI” responses are removed, this increases to 52%. This puts the “No, I’m not concerned” response marginally ahead of those related to concerns about AI use:
However, the highest of these concern areas is related to the limited use of AI. When this 21% is added to the 52%, close to three-quarters of end-user respondents can be considered to be “pro AI”.
This is the only survey question that was asked to both IT would you like AI to not penetrate? professionals and end-users.
The top three most-stated areas were:
The comparison of end-user and IT professional would you like AI to not penetrate? responses found them to be similar as shown in the table. End-users were, on the whole, slightly more against AI use than IT professionals, including in two of the shared top three areas.
End-users were also more concerned about using AI for handling customer data, talent acquisition, performance management, and quality control and accountability.
11% of respondents were happy for AI to penetrate all of the potentially problematic area options. That’s more than double the percentage of IT professionals.
We believe that the two discrete data sets targeting IT professionals and end users do offer valuable and holistic insights into the usage of AI. While most organizations are still in the early stages of AI adoption, IT leaders do realize the tremendous benefits of adopting AI in IT. Overcoming the common barriers to using AI in IT operations, such as AI inconsistencies or security adherence, can help companies bridge the gap between the estimated value of AI to realizing its actual benefits in their IT environment.