Automation is transferring 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 onerous 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 of us on the market, it is so simple as ABC. Nevertheless, many technical leads and CIOs discover themselves in hassle on the beginning line,” says Muhammad Nabeel, chief expertise officer at Start, an leisure streaming service in Pakistan.
At problem from the beginning are the same old firm politics and AI — which could 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 expertise. Due to this fact, any CIO or head of expertise 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 high of which can be all of the AI issues, too.
“This 12 months, hidden prices and regulatory curveballs will chunk if ignored. Past licensing charges, look ahead to integration spaghetti — methods that don’t “discuss” easily — and coaching gaps that stall adoption. New information privateness rules, 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 advised, there’s rather a lot for IT to handle unexpectedly. For the sake of sanity and technique, maybe it is best to first take into account the pitfalls and challenges earlier than making an attempt to map out a method.
Pitfall 1: Operating into obstacles you’ll be able to’t see
Within the strategy of implementing automation and getting all of the transferring elements proper, generally individuals neglect 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 intention of automating processes is to enhance how the enterprise performs. Which means 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 expertise. CIOs want to make sure that they will translate AI capabilities into concrete enterprise methods to show robust ROI potential for stakeholders,” says Eric Johnson, CIO at PagerDuty, an AI-first operations platform.
Pitfall 2: Underestimating information high quality points
Information is arguably probably the most boring problem 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 information and the job is rarely fairly accomplished. It might be boring work, however it’s important and could be deadly if left for later.
“Some of the vital errors CIOs make when approaching automation is underestimating the significance of information high quality. Automation instruments are designed to course of and analyze information at scale, however they rely fully on the standard of the enter information,” says Shuai Guan, co-founder and CEO at Thunderbit, an AI internet scraper instrument.
“If the info is incomplete, inconsistent, or inaccurate, automation is not going to solely fail to ship significant outcomes however can also exacerbate present points. For instance, flawed buyer information fed into an automatic advertising and marketing system might result in incorrect focusing on, wasted sources, and even reputational injury,” Guan provides.
Pitfall 3: Mistaking the duty for the aim
A typical method is to automate the straightforward, repetitive processes with out giving thought to an issue that lurks beneath. Ignoring or overlooking the trigger now might show extremely damaging ultimately.
“CIOs usually fall into the entice of pondering automation is nearly suppressing noise and lowering ticket volumes. Whereas that’s one pretty widespread use case, automation can provide rather more worth when accomplished 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 greater dangers down the street. It’s usually the suppressed Severity 3-4 problem that when left unattended, turns into the S1 or 2 extra time!” Gaston says.
Keep in mind additionally that enterprise targets and applied sciences change over time and so too should processes.
“Deal with high-impact areas, leverage the facility of open-source instruments initially, and monitor the end result. Change when and the place vital. Don’t undertake the ‘fireplace and neglect” 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 effectively.
“One mistake is assuming AI-driven automation can run with out human oversight. AI is a robust instrument, nevertheless 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.
Nevertheless, even conventional automation instruments require integration. Most in IT are conscious of this nevertheless it doesn’t imply that planning for it made it into the ultimate technique.
“One other problem is failing to plan for integration, which may result in vendor lock-in and disconnected methods. 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 slightly a rush to automate the low-hanging fruit to indicate a quick win. Sadly, a quick win isn’t essentially the identical as a giant win. A value-benefit evaluation will steer you true whereas a fast choose may lead you astray.
“For pipelines that happen much less regularly or require little time, automation gives lesser worth. Like most enterprise processes, a value could 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 solution to save effort, time and wasted value.
“Any course of that requires complicated human reasoning, emotion, or interplay, or doesn’t observe established guidelines and constructions, usually are not appropriate for automation. In fact, AI is blurring that distinction and getting higher at simulating complicated human behaviors and establishing constructions the place none appear to exist. Nevertheless, contemplating the present state of improvement and potential 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 research, 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 doubtless that the ROI gained’t be constructive earlier than the method itself turns into out of date,” Biswas provides.
Pitfall 6: Focusing solely on value
Provided that economies are unsure all over the world from inflation, political upheaval, and different elements, it’s comprehensible that value considerations are elevated now. However that slim focus can depart you blind to different price range impacts.
“CIOs threat selecting the fallacious expertise, resulting in integration challenges, pointless complexity, or vendor lock-in. A typical pitfall is focusing solely on value financial savings slightly than broader advantages like agility, innovation, and buyer expertise, which may restrict the precise worth of automation,” says Derek Ashmore, software transformation principal at Asperitas, an IT consultancy.
Rising New Challenges
2025 is ushering in quite a lot 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 12 months, CIOs must be notably vigilant about rising regulatory necessities that would influence their automation methods. Staying knowledgeable about industry-specific rules and compliance requirements is important, particularly relating to how automated methods deal with information,” says Chris Drumgoole, EVP, World Infrastructure Companies at DXC Know-how, a world expertise companies supplier.
It isn’t simply federal rules you could watch intently, however regional and state rules too.
“The combination of AI into IT automation is accelerating, with applied sciences like generative AI and agentic AI taking part in pivotal roles. State legislatures within the US are actively introducing AI-related payments, with a whole lot proposed in 2025,” says Ashmore.
Ashmore warns that these legislative efforts embody complete shopper safety, sector-specific rules on automated decision-making, chatbot oversight, generative AI transparency, information heart 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 connected to automation implementations.
Surprising bills in operationalizing AI, rising complexity in multi-cloud integration, and integration necessities throughout rising ecosystems are all placing stress on IT, in accordance with Deepak Singh, president and chief expertise officer at Adeptia, an AI and self-service platform.
Additionally lurking within the background, however quickly to lift 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 accomplished with out company oversight or interference. On high of that’s the rising variety of AI fashions built-in or embedded in enterprise software program and {hardware}, in addition to in personal units like smartphones. That’s quite a lot of unattended and probably unsecured AI wandering round within the group. For instance, that’s quite a lot 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 all types.
“Leaders ought to concentrate on upskilling the expertise they have already got and investing in communities to construct robust expertise pipelines. This manner, as automation will increase, the workforce can take an oversight function and luxuriate in extra capability to concentrate on 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 consultants 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 guaranteeing expertise can span the manufacturing and IT disciplines. This will take the type of educating manufacturing employees in IT however should additionally concentrate on guaranteeing that IT employees and companions perceive the actual challenges and information surroundings on the manufacturing ground. IT and OT (Operational Know-how) can not have partitions between them and should function towards widespread targets with adequate understanding of one another’s domains,” Gaus says.
