Enterprises are being stung by unexpected AI bills, and it’s disrupting IT budgets
A survey by Cisco found that observability costs are skyrocketing amid the shift to AI, despite half unable to measure LLM response quality
More than half of UK organizations have reported "surprise" bills for monitoring IT systems, according to new research from Cisco, and AI is a key factor behind surging costs.
In a new survey from the firm, 65% of respondents said their observability costs had exceeded infrastructure or operational costs this year, with half reporting surprise costs with telemetry data, and another third saying that led to reduced data coverage or retention.
The complaint isn't just high costs, but those that are higher than expected, with 55% of those asked pointing to surprise expenses that are straining operations more than typically baseline costs.
General AI adoption costs are also exacerbating the issue, Cisco found. Another 55% said they are finding it impossible to control AI costs, a problem as AI observability has become a priority, according to nine-in-ten respondents.
That echoes a previous study by KPMG, that found many companies struggle with costs and lack visibility into AI spending.
“For today’s enterprises, the excitement of AI is clashing with the reality of the bill,” added Kamal Hathi, SVP and GM, Splunk at Cisco.
It's not just Cisco spotting this trend, either. A poll by Asana earlier this year found that eight-in-ten UK IT leaders had seen unexpected costs increases from AI over the last year, with Gartner predicting that AI costs could exceed developer salaries by 2028.
Sign up today and you will receive a free copy of our Future Focus 2026 report - the leading resource for IT decision-maker insight on priorities and investment areas in AI, security and more.
Given the rising costs, companies began cracking down on "tokenmaxxing" and other AI overuse this summer, with Uber introducing a monthly AI cap and Accenture telling staff to cut back on AI use for basic tasks.
AI observability
While monitoring AI is a clear priority, more than half of those polled said they don't have a clear business outcome tied to their AI usage, and 56% said they're unable to measure AI workload metrics such as LLM response quality, which Cisco said was a core metric.
Indeed, 85% said they need to have visibility into AI hallucinations, but fewer than six-in-ten said they're effective at spotting such concerns.
“We believe observability is the essential bridge to restoring predictability and confidence to the agentic AI era," Hathi added.
"By providing the necessary guardrails and incorporating observability into an AI tokenomics framework, leaders can transform AI from an unpredictable expense into a measurable and strategic long-term business investment."
Collaboration and teamwork
One silver lining in the confluence of challenges faced by enterprises lies in heightened teamwork and collaboration, according to Cisco. More than half (55%) of those polled are worried about the unintended consequences of AI agents, prompting closer synergy between departments.
Eight-in-ten IT professionals have observed a marked increase in collaboration between observability and security teams as a result.
Nearly three-quarters (73%) also revealed they can accurately trace performance issues to security root causes, an increase from 65% in the survey last year. Similarly, there’s been a marked increase in data sharing between these two teams from 74% to 84% over the last year, with a similar leap seen in working together on troubleshooting.
"Most companies will soon have far more agents at work than people," said Jeetu Patel, president and chief product officer (CPO) at Cisco.
"Managing agents at scale requires reimagining both security and observability. Have our agents been compromised? Are they doing their job well? In the world of agentic AI, these questions are inseparable, and they're moving at machine speed."
FOLLOW US ON SOCIAL MEDIA
Follow ITPro on Google News and add us as a preferred source to keep tabs on all our latest news, analysis, views, and reviews.
You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.
Freelance journalist Nicole Kobie first started writing for ITPro in 2007, with bylines in New Scientist, Wired, PC Pro and many more.
Nicole the author of a book about the history of technology, The Long History of the Future.
-
CIOs need to take back control of their IT roadmapIn-Depth IT leaders are facing difficult decisions about their ERP estates – but experts say taking an ‘AI last’ approach can avoid costly disruption
-
TfL takes Oyster photocard systems offline following breachNews Security experts warn the Oyster photocard breach could lead to follow-up attacks on victims
-
A big tech taskforce has unveiled a blueprint for UK AI leadership – but time for action is running outNews Success depends on deeper collaboration between government, industry, local leaders and communities
-
Cisco just launched two cyber-focused small language models: Antares-350M and Antares-1B aim to supercharge codebase analysis – and they run at a “fraction of the compute expense” of popular frontier modelsNews The Antares models unveiled by Cisco aim to cut costs in codebase analysis
-
‘Cisco now is delivering that critical infrastructure for the AI era’: Cisco's infrastructure unification push aims to simplify management for the agentic eraNews The company aims to put the power in customers’ hands while emphasizing the importance of network efficiency
-
Cisco teams up with DSIT to drive digital skills adoptionNews Partnership supports the government's TechFirst program to provide one million secondary school students with access to digital learning experiences
-
Cisco Wireless CTO: Mastering connectivity is the key to driving AI success and enterprise productivity – but beware of the ‘wireless AI paradox’News Enterprises are ramping up wireless connectivity investment as AI-related network demands grow
-
CISOs are keen on agentic AI, but they’re not going all-in yetNews Many security leaders face acute talent shortages and are looking to upskill workers
-
AI is “forcing a fundamental shift” in data privacy and governanceNews Organizations are working to define and establish the governance structures they need to manage AI responsibly at scale – and budgets are going up
-
Cisco looks to showcase “unique value” with revamped 360 Partner ProgramNews Cisco has unveiled a revamped partner framework to help partners capitalize on growing AI-driven customer demand