New Cisco AI study sees widening execution gap, strain on manufacturing infrastructure

A brand-new State of Industrial AI Report from Cisco shows AI deployment accelerating but placing significant strain on manufacturing IT and OT infrastructure that must—as the findings from 350 manufacturing decision-makers worldwide also note—be modernized to support the surging technology.

The results of the new research from the Silicon Valley-based networking hardware, software, and telecom equipment giant deal much with AI, but they also reveal much about the state of manufacturing infrastructure required to support artificial intelligence deployments.

See also: Top-performing OEMs likelier to use tech such as digital twins, cobots to cut downtime, speed recovery, Rockwell finds

Cisco surveyed decision-makers in 19 countries, with annual revenues of over $100 million each, and undertook the work with London-based B2B-focused Sapio Research.

Some of the notable findings—particularly about network infrastructure and how industrial AI demands modernization—from the Cisco report show how interconnected the two are:

• 56% said unreliable wireless connectivity frequently disrupts manufacturing operations.
• Of the respondents, almost all—96%—agreed wireless connectivity is critical to AI success, and a majority—51%—anticipate significant increases in connectivity and reliability requirements.
• Reliable connectivity (49%), edge compute (44%), and bandwidth (39%) are top network requirements for AI to succeed at scale—meaning that software, infrastructure, or services connected to the technology can handle high volumes of traffic, data, or users automatically, without performance degradation.

The report did note how quickly AI is being deployed in manufacturing environments, however Cisco found qualifiers that could temper exuberance about the technology:

• 68% of manufacturing organizations are already deploying AI, yet only 19% consider their deployments mature, highlighting a widening execution gap. This points to another finding: That most organizations are deploying, but most aren’t doing so at scale because of infrastructure and other concerns.
• However, 80% of manufacturing leaders did say that failing to invest in AI risks falling behind competitors, which reflects the pressure they’re under to adopt, despite the limitations of their existing technology and infrastructure.
• Process automation (66%) and automated quality inspection (54%) are the most widely deployed use cases for AI in manufacturing.

Security constraints shape deployment decisions

The Cisco State of Industrial AI Report also has much to say about cybersecurity and how it also remains a top barrier to deploying AI at scale.

In the Cisco/Sapio research, 40% of the respondents identified cybersecurity as the top obstacle to scaling AI, particularly around issues such as segmentation, data exposure, and ransomware risk, while 85% expect AI to improve detection of cyber-threats and resilience.

See also: Workshop confronts manufacturing execs with the big stakes that ride on proper cybersecurity protocols

Also, while 54% of respondents expected returns on their AI investments within a year, that places pressure on platforms that can support fast deployment but can compromise stability or security.

Among the respondents, 81% expect AI to improve their cybersecurity posture. While security gaps are limiting AI scale today, manufacturers view AI as a tool to strengthen detection, monitoring, and resilience.

“Cybersecurity concerns are significantly limiting AI adoption by creating a ‘trust deficit’ and introducing new, complex risks that outpace traditional security measures,” according to the Cisco report. “A recent Forrester report cited that, 'among AI decision-makers, 29% identify trust as the single largest barrier to generative AI adoption in their organizations.’”

“Manufacturers see the greatest benefit where AI and cybersecurity are designed together, rather than treated as separate initiatives.”

IT/OT collaboration emerges as a technical, organizational necessity

Cisco noted that collaboration between IT and OT teams in a manufacturing operation is key to deploying AI at scale. But a large segment of respondents in the Cisco/Sapio research also reported little to no collaboration between IT and OT at their organizations. Specifically:

• 43% of the surveyed manufacturing organizations have little to no IT/OT collaboration.
• 34% cite lack of collaboration between IT and OT teams as a major challenge that is limiting AI-enabled operations at their organizations.
• 28% say OT domain expertise is critical to scaling AI.

Organizations that reported stronger collaboration between IT and operational teams also report higher confidence in scaling AI and fewer network and security issues, according to Cisco.