The State of AI in ITSM 2025

Last December, Atomicwork, PeopleCert, and ITSM.tools published a State of AI in IT 2025 report. This followed an early 2024 IT service management (ITSM)-focused report on artificial intelligence (AI) adoption in North American IT organizations. This time, the IT element of the survey is global, thanks to the assistance of PeopleCert.

As with the earlier report, there were two surveys. The first survey posed AI-related questions to 346 IT professionals and the second to 700 end-users. The IT and end-user surveys were conducted in Q4 2024 to provide a service provider and service receiver view for 2025.

This article shares some of the insights from the new data (and has links to our latest quick poll – please help). The full report (but this doesn’t include correlations) is available to download here.

An interesting observation on trust

Notwithstanding that “correlation does not imply causation,” there was an interesting set of correlations in the IT professional survey data that I summed up in the following statement:

1. AI-focused budgets are important. For example, most Negative AI ROI responses (94%) came from organizations with No allocation or less than 10% of the IT budget allocated to IT initiatives. Conversely, 71% of Positive AI ROI and only 19% of Neutral responses came from organizations with more than 10% of the IT budget spent on AI initiatives.
2. C-suite-led AI initiatives struggle when compared to those originated by IT. For example, the AI investments originating from the C-suite were least likely to have a positive ROI and significantly more likely to have a negative ROI.
3. Trust in AI-driven IT systems correlates with AI ROI success – all of the Negative AI trust responses were from respondents reporting a negative ROI from AI investments.

There appears to be a “chain” here – that AI investment leads to AI success, which leads to AI trust (or it could be that AI trust leads to AI spending, which leads to AI success). It will be interesting to see whether this perceived “chain” has merit when further surveys are conducted.

Some key AI in ITSM 2025 survey data points – IT professionals

• IT leadership (34%) and C-suite (23%) were the highest originators of AI initiatives. This C-suite level is similar to the previous report’s 24%. When the Not applicable, Don’t know, and Other responses are removed, this is IT leadership 47% and C-suite 31%, with the IT team accounting for the remaining 22%.
• 41% of respondents stated that their organizations have an IT budget allocation for AI initiatives. When the Don’t know and Not applicable responses are removed, this is 71%. However, IT budget allocation to AI initiatives seems low – 16% No allocation and 21% Less than 10%. This was 64% spending less than 10% of their IT budget on AI initiatives and 36% more than 10% when the Not applicable and Don’t know responses are removed. As with the 2024 data, IT companies influence the data and results. Unsurprisingly, software companies make up close to 50% of all companies spending more than 20% of their IT budgets on AI. The next “big spender” was Healthcare (albeit with a smaller sample size).
• The top areas where AI has impacted ITSM were data analysis (55%), end-user assistants (48%), knowledge management (43%), and incident management (39%).
• The top realized benefits were increasing employee productivity (40%), improving user experience (33%), optimizing operations and reducing costs (29%), and enabling better decision-making (28%).
• The current top concerns or barriers to AI adoption were governance and compliance (51%), customer data security (47%), and employee data security (43%); lack of expertise and costs were at 41% and 35%, respectively. This had changed since the previous report, where the top concerns or barriers were customer data security (42%), additional cost (39%), and inaccuracy or inconsistency (33%).
• 55% of respondents don’t trust AI to make decisions without human oversight, versus 25% that do. When Don’t know and Other responses are removed, this is 31% trust and 69% distrust. Although, 47% of respondents trusted AI more than they did 12 months ago. While 12% trust it less, and 40% have stayed the same (which could be trust or distrust).
• Responsible and ethical AI (79%) and security-focused (business level) (79%) were considered the most important factors in AI implementation; oddly, user-centric was in last place (51%).
• The top areas for AI not to enter were ethical and legal decision-making (51%), people management (42%), and handling sensitive business data (38%). However, 35% of respondents thought AI would eventually be used in all areas of their business.

Some key AI in ITSM 2025 survey data points – end-users

• Only 16% of survey respondents hadn’t contacted their corporate IT support in the last 12 months. When the Unsure and We don’t have an IT support team responses are removed, this is still only 18%.
• 42% of survey respondents stated that their organizations use AI. Of these, 78% thought the AI helped, while 22% didn’t. The story differed for the respondents without AI, though – of the 38% without AI, 58% were happy there was no AI, and 42% wanted there to be. When the “Don’t know” responses are removed, 53% of survey respondents stated that their organizations use AI, compared to 47% who don’t.
• 71% of survey respondents use ChatGPT or similar free AI tools for their work, with roughly half of the full sample (49%) using it at least weekly. This is 70% of the respondents using free AI tools (71%) using it at least weekly. Interestingly, there was a strong correlation between respondents who use ChatGPT (and similar) and whether their IT team uses AI for support. Conversely, those respondents without corporate AI use were less likely to use ChatGPT and other free AI tools for their work.
• 44% of survey respondents weren’t concerned about their organizations using AI. When the Don’t know and We’re not using AI responses are removed, this is 60%. 30% were concerned, though, split across: using AI too liberally (12%), the use of AI is limited (14%), and it’s hindering and not helping (4%).
• 31% of survey respondents trusted AI-driven IT systems to make autonomous decisions without human oversight versus 44% who didn’t. When the Don’t know and It depends responses are removed, this was 41% and 59%. 64% of survey respondents trust AI systems more than 12 months ago, while 36% trust them less. However, there was no unchanged option as in the IT survey (where 47% trust AI more and only 12% trust AI less).

We hope that this early AI in ITSM 2025 snapshot is helpful. The full report can be downloaded from the Atomicwork website.