Automation is shifting from a routine IT job to a race to cross an ill-defined end line. AI tends to be the bug smearing the windshield and making it laborious to see the place you’re headed. Highway hazards are additional complicating the drive to elevated effectivity.
«For some, automation is a buzzword and an uphill battle, however for many technical people on the market, it is so simple as ABC. Nonetheless, many technical leads and CIOs discover themselves in hassle on the beginning line,” says Muhammad Nabeel, chief know-how officer at Begin, an leisure streaming service in Pakistan.
At difficulty from the beginning are the same old firm politics and AI — which will be harder to barter than bean counters and C-suite heavyweights mixed.
“These days, AI has a drastic affect on each stroll of life, particularly know-how. Subsequently, any CIO or head of know-how should incorporate the AI issue,” Nabeel provides.
Though AI is a dominant power, it isn’t the one play in automation. Some established instruments and guidelines nonetheless apply. Sadly, so do the earlier pitfalls and challenges. Heaped on prime of which might be all of the AI issues, too.
“This yr, hidden prices and regulatory curveballs will chew if ignored. Past licensing charges, look ahead to integration spaghetti — programs that don’t “discuss” easily — and coaching gaps that stall adoption. New information privateness laws, like evolving GDPR [the European Union’s General Data Privacy Regulation] and AI transparency legal guidelines, imply CIOs should vet instruments for compliance and moral design,” says Dawson Whitfield, CEO and co-founder of Looka, an AI platform for designing logos.
All instructed, there’s rather a lot for IT to handle . For the sake of sanity and technique, maybe it is best to first think about the pitfalls and challenges earlier than attempting to map out a technique.
Pitfall 1: Working into obstacles you’ll be able to’t see
Within the strategy of implementing automation and getting all of the shifting elements proper, generally individuals overlook to first consider the method they’re automating.
“You don’t know what you don’t know and may’t enhance what you’ll be able to’t see. With out course of visibility, automation efforts might result in automating flawed processes. In impact, accelerating issues whereas losing each time and sources and resulting in diminished goodwill by skeptics,” says Kerry Brown, transformation evangelist at Celonis, a course of mining and course of intelligence supplier.
The goal of automating processes is to enhance how the enterprise performs. Meaning drawing a direct line from the automation effort to a well-defined ROI.
«When evaluating AI and automation alternatives for the group, there are sometimes gaps in understanding the enterprise implications past simply the know-how. CIOs want to make sure that they will translate AI capabilities into concrete enterprise methods to display robust ROI potential for stakeholders,” says Eric Johnson, CIO at PagerDuty, an AI-first operations platform.
Pitfall 2: Underestimating information high quality points
Knowledge is arguably essentially the most boring difficulty on IT’s plate. That’s as a result of it requires a ton of effort to replace, label, handle and retailer large quantities of knowledge and the job is rarely fairly achieved. It could be boring work, however it’s important and will be deadly if left for later.
“One of the vital vital errors CIOs make when approaching automation is underestimating the significance of knowledge high quality. Automation instruments are designed to course of and analyze information at scale, however they rely solely on the standard of the enter information,” says Shuai Guan, co-founder and CEO at Thunderbit, an AI internet scraper software.
“If the info is incomplete, inconsistent, or inaccurate, automation is not going to solely fail to ship significant outcomes however may additionally exacerbate present points. For instance, flawed buyer information fed into an automatic advertising system may result in incorrect concentrating on, wasted sources, and even reputational injury,” Guan provides.
Pitfall 3: Mistaking the duty for the aim
A typical method is to automate the simple, repetitive processes with out giving thought to an issue that lurks beneath. Ignoring or overlooking the trigger now might show extremely damaging in the long run.
«CIOs typically fall into the entice of pondering automation is nearly suppressing noise and decreasing ticket volumes. Whereas that’s one pretty widespread use case, automation can supply rather more worth when achieved strategically,” says Erik Gaston, CIO of Tanium, an autonomous endpoint administration and safety platform.
“If CIOs focus solely on suppressing low-level tickets with out addressing the basis causes or understanding the broader patterns, they threat permitting these points to snowball into extra extreme issues that may finally result in larger dangers down the highway. It’s typically the suppressed Severity 3-4 difficulty that when left unattended, turns into the S1 or 2 time beyond regulation!” Gaston says.
Keep in mind additionally that enterprise objectives and applied sciences change over time and so too should processes.
“Concentrate on high-impact areas, leverage the ability of open-source instruments initially, and monitor the result. Change when and the place essential. Don’t undertake the ‘hearth and overlook» precept,’» says Nabeel.
Pitfall 4: Failing to plan for integration
Integration turns into a necessity in some unspecified time in the future. With AI, integrating with human overseers is a right away want. Usually it have to be built-in with different software program as properly.
“One mistake is assuming AI-driven automation can run with out human oversight. AI is a strong software, however it nonetheless requires human checks to catch errors, bias, or safety dangers,” says Mason Goshorn, senior safety options engineer at Blink Ops, an AI-powered cybersecurity automation platform.
Nonetheless, even conventional automation instruments require integration. Most in IT are conscious of this however it doesn’t imply that planning for it made it into the ultimate technique.
“One other problem is failing to plan for integration, which might result in vendor lock-in and disconnected programs. CIOs ought to select automation instruments that work with present infrastructure and assist open requirements to keep away from being trapped in a single supplier’s ecosystem,” says Goshorn.
Pitfall 5: Not permitting the info to drive choices in what to automate
Usually the plan isn’t actually a plan however relatively a rush to automate the low-hanging fruit to indicate a quick win. Sadly, a quick win isn’t essentially the identical as an enormous win. A value-benefit evaluation will steer you true whereas a fast choose would possibly lead you astray.
“For pipelines that happen much less regularly or require little time, automation supplies lesser worth. Like most enterprise processes, a price will be related to automation, and the price financial savings ought to exceed the price of implementation and upkeep,” says David Brauchler, technical director & head of AI and ML safety at cybersecurity consultancy, NCC Group, a cybersecurity firm.
Figuring out what processes shouldn’t be automated early on is one other method to save effort, time and wasted value.
“Any course of that requires advanced human reasoning, emotion, or interplay, or doesn’t comply with established guidelines and buildings, should not appropriate for automation. In fact, AI is blurring that distinction and getting higher at simulating advanced human behaviors and establishing buildings the place none appear to exist. Nonetheless, contemplating the present state of improvement and attainable authorized and ethical ramifications, such processes must be deprioritized for automation,” says Sourya Biswas, technical director, threat administration and governance, NCC Group.
“Additionally, contemplating the lead time to investigate, implement and combine automation, any course of topic to main adjustments in working situations within the close to future shouldn’t be thought of for automation as it’s probably that the ROI gained’t be optimistic earlier than the method itself turns into out of date,” Biswas provides.
Pitfall 6: Focusing solely on value
Provided that economies are unsure world wide from inflation, political upheaval, and different elements, it’s comprehensible that value issues are elevated now. However that slim focus can go away you blind to different finances impacts.
“CIOs threat selecting the fallacious know-how, resulting in integration challenges, pointless complexity, or vendor lock-in. A standard pitfall is focusing solely on value financial savings relatively than broader advantages like agility, innovation, and buyer expertise, which might restrict the precise worth of automation,” says Derek Ashmore, utility transformation principal at Asperitas, an IT consultancy.
Rising New Challenges
2025 is ushering in plenty of new challenges for IT to surmount in automation implementations. Though adjustments in regulation and related compliance prices are ongoing points, they’re much more so now.
“This yr, CIOs must be notably vigilant about rising regulatory necessities that would impression their automation methods. Staying knowledgeable about industry-specific laws and compliance requirements is crucial, particularly relating to how automated programs deal with information,” says Chris Drumgoole, EVP, International Infrastructure Companies at DXC Know-how, a worldwide know-how providers supplier.
It isn’t simply federal laws you could watch intently, however regional and state laws too.
“The combination of AI into IT automation is accelerating, with applied sciences like generative AI and agentic AI enjoying pivotal roles. State legislatures within the US are actively introducing AI-related payments, with tons of proposed in 2025,” says Ashmore.
Ashmore warns that these legislative efforts embrace complete shopper safety, sector-specific laws on automated decision-making, chatbot oversight, generative AI transparency, information middle power utilization, and public security regarding superior AI fashions.
“This surge in state-level regulation provides complexity to compliance for organizations implementing IT automation,” Ashmore provides.
Among the rising challenges are extra immediately hooked up to automation implementations.
Sudden bills in operationalizing AI, rising complexity in multi-cloud integration, and integration necessities throughout rising ecosystems are all placing strain on IT, in response to Deepak Singh, president and chief know-how officer at Adeptia, an AI and self-service platform.
Additionally lurking within the background, however quickly to boost its ugly head, is the issue of a rising shadow AI. Enterprise customers are routinely turning to free and low-cost AI subscription fashions to get their work achieved with out company oversight or interference. On prime of that’s the rising variety of AI fashions built-in or embedded in enterprise software program and {hardware}, in addition to in non-public gadgets like smartphones. That’s plenty of unattended and probably unsecured AI wandering round within the group. For instance, that’s plenty of AI that may be gathering information to coach future AI fashions on, and a few of that information could also be proprietary.
Final, however definitely not least, is the dearth of expertise essential to remake enterprise processes in AI’s picture and match for automation instruments of every kind.
“Leaders ought to deal with upskilling the expertise they have already got and investing in communities to construct robust expertise pipelines. This fashion, as automation will increase, the workforce can take an oversight function and luxuriate in extra capability to deal with innovation that may enhance the underside line,” mentioned Tim Gaus, Good Manufacturing enterprise chief at Deloitte, a consulting agency.
The important thing to success lies in coaching area specialists to make use of AI and different applied sciences, and to precisely consider what processes can and may’t be efficiently automated.
“Key to this for producers is making certain expertise can span the manufacturing and IT disciplines. This could take the type of educating manufacturing employees in IT however should additionally deal with making certain that IT employees and companions perceive the actual challenges and information setting on the manufacturing flooring. IT and OT (Operational Know-how) can not have partitions between them and should function towards widespread objectives with adequate understanding of one another’s domains,” Gaus says.