The promise that artificial intelligence will streamline IT operations through automation may be coming true in some instances, but in others AI is adding to the complexity of the IT environment, placing departments under pressure, exposing fragmentation and increasing cost.
This is according to Thomas Meyer, vice president for research at market intelligence firm International Data Corporation (IDC), who spoke at the Huawei Cloud Summit in Barcelona, Spain on Sunday ahead of Mobile World Congress this week.
“One part has to do with deciding what to do with all the new kinds of technologies that are coming up. The other piece is you have to connect all the dots: the computing capacity that you have, the hardware capex cost and the money that you will have to continuously spend going forward, cloud instances, latency issues and the LLMs (large language models) you use,” said Meyer.
“The challenges are around the technical debt that is being created because it is tough to keep up with innovation.”
Meyer said another factor putting IT leaders and their departments under pressure is the fragmentation of organisational data. AI workloads typically work best when the data they use is extracted from a “single source of truth”. Taking advantage of AI in organisations is therefore often hampered by complex integrations that must take place beforehand. According to Meyer, a well-organised “data estate” is key to AI readiness.
Data maturity
IDC’s recommended approach involves four levels of data maturity, starting with basic inventory and source identification, progressing through intelligence engineering and governance, into analytics and AI training, and finally arriving at optimisation and action.
Most organisations, Meyer said, are tempted to skip levels and, as a result, they struggle.
“With the fast pace of innovation, it is tough keeping up with the integration complexities. On the data side, the house needs to be put in order and it is not going to happen by itself,” said Meyer.
Read: Vibe coding is transforming development – but at what cost to open source?
Alongside the pressure to keep up with the latest technologies, IT leaders are also expected to show a return on investment on AI spend.
IDC data shows that, on average, large organisations run 30 to 40 AI proof-of-concepts before any meaningful scale is achieved. Of those, roughly five make into production, with only two of three of those becoming genuinely successful. Meyer said only 11% of organisations have AI implementations that are truly successful and suggests taking a longer-term view on AI investments.
“A lot of companies happily move along using traditional ROI (return on investement) metrics. But AI is fundamental, just like the internet, just like cloud and other things that were really disruptive. You need to be in it for the long term and you need to describe your metrics and expectations along those lines,” said Meyer.
One of the main pressure drivers is the understanding that AI is changing how business is fundamentally done, forcing processes, business models, partners and even customers to change. According to Meyer, some of the transformative potential associated with the technology and putting IT departments under pressure is pure hype, especially from vendors trying to sell their particular solutions.
Read: How AI is rewriting the rules of consulting
“I don’t believe you need to believe all the hype that is in the market, particularly on the vendor side. But I think everyone in this room understands that this is a reality that will come. It may be slower, but it will come and we need to get ready for it,” said Meyer. — (c) 2026 NewsCentral Media
Get breaking news from TechCentral on WhatsApp. Sign up here.
