Telecoms warned against rushing AI rollouts, survey finds

Designit has published research suggesting telecom companies may be moving too quickly when deploying artificial intelligence, with speed of rollout identified as the sector's biggest mistake.

The survey found 43% of global design and telecom professionals believe deploying AI too fast is the main error operators make when introducing the technology. Treating AI as a cost-cutting tool ranked second at 32%, followed by poor data readiness at 14% and automating without organisational redesign at 11%.

The findings highlight a tension in the telecom industry's approach to AI. Operators face pressure to show quick results, but many businesses may not be giving the same attention to the organisational changes needed to support wider adoption.

Just 11% of respondents identified organisational redesign as a priority, despite the view that AI often requires changes to workflows, governance and accountability if it is to move beyond isolated pilots. The research suggests that gap helps explain why some projects fail to meet early expectations.

Telecom companies have been among the most active large industries in testing AI, using it in customer service, network operations and internal processes. Yet the sector also faces some of the toughest operating conditions, with ageing infrastructure, tight regulatory oversight and large customer bases that leave little room for disruption when new systems are introduced.

Jakob Voldum, Insights and Strategy Design Director at Designit, said: "Telecoms is one of the most complex environments in which to embed AI. Managing legacy infrastructure, regulatory pressure and the need to serve millions of customers every day means the margin for a poorly prepared AI deployment is slim.

"In many cases, organisations are introducing AI without a clear view of what success looks like. Without that direction, speed only brings misalignment to the surface quicker.

"Technology is rarely the problem in isolation. The challenge is ensuring the organisation is ready to meet it: redesigning workflows, clarifying accountabilities, and putting governance and team structures in place before the pressure to scale arrives.

"Those who succeed will treat AI as an opportunity to redesign how they operate, not just where they automate. That is where long-term value will be created."

Operational gap

The research argues that AI progress in telecoms is still often judged by visible measures such as how quickly a tool goes live or how many processes it touches. That emphasis can make deployment look successful in the early stages, even when a business has not prepared its operating model for broader use.

In practice, this can leave pilots stranded between experimentation and adoption. A tool may work technically, but still struggle to fit existing decision-making structures, team responsibilities or service processes.

For telecom operators, the issue matters because AI projects usually cut across multiple departments. Network teams, customer operations, compliance functions and commercial units may all need to work together if a system is to be integrated into day-to-day operations rather than remain a trial.

Designit cited its work with Dutch telecom provider Odido as an example of a more integrated approach. The company said the work involved an embedded studio model that placed teams alongside Odido's business and IT functions to align commercial aims with operational delivery.

The approach was presented as a way to move AI from a series of individual projects to a cross-functional part of the business. The research did not provide financial figures or deployment targets linked to that work.

Industry pressure

The telecom sector has faced sustained pressure to modernise while controlling costs, a combination that has made AI particularly attractive to executives seeking efficiency gains and service improvements. But the survey results indicate that an emphasis on speed can distort priorities when internal structures are not ready for the change.

Respondents also highlighted the risk of viewing AI mainly as a cost-cutting measure. That concern, which ranked second in the study, suggests some industry professionals believe companies may be narrowing AI's role before understanding where it fits operationally.

Poor data readiness also remained a notable issue. While fewer respondents selected it as the main mistake, data quality and access have long been central barriers to AI adoption in large telecom organisations that rely on fragmented legacy systems.

Taken together, the findings suggest the challenge for operators is not simply whether to deploy AI, but whether they can reshape their organisations to use it effectively at scale. As Voldum put it, "Those who succeed will treat AI as an opportunity to redesign how they operate, not just where they automate. That is where long-term value will be created."