Cisco report: trust, not tech, limits AI scaling

The report arrives at a moment when the enterprise AI conversation has spent years centered on capability — which model is smartest, which GPU is fastest — a framing the body says made sense when technology was the constraint. Cisco's argument is that this is no longer the case: AI can now do far more than most organizations will allow, and the gap is trust. The Australian sample is presented as a deliberately hard test case, since the cited University of Melbourne and KPMG research puts Australian trust in AI at 36%, the lowest among 47 countries surveyed, with only 30% believing its benefits outweigh its risks. Cisco ANZ CTO Carl Solder frames the upside of solving it there: the resulting playbook "travels well beyond Australia," per Stefan Leitl of Cisco Australia and New Zealand.

The panel discussion points to culture as the pillar organizations most often skip. Solder described a customer that invested in infrastructure, governance, data and IT skills but not culture, and met "a big pushback from their employees and customers" on the toolsets it had bought. Flight Centre CTO Chris Locke offered the counter-example of the company's 77-year-old founder using Anthropic's Claude daily, while TBWA Australia chief AI officer Lucio Ribeiro described trust as built on ownership, competence, consistency and transparency. Funlab's Naz Wilson said early "AI Super Squad" sessions mainly clarified which tools were safe, because many employees simply did not know what they were allowed to do.

The body also flags the opposite risk — over-trust. Elmo Software's Anna Tosky called out "AI slop" reports sent without review, and argued the fix starts with redesigning the process rather than throwing agents at it. On agents that act rather than draft, fewer than half of respondents are very or extremely confident their foundations are ready, and Cisco's 2026 Relentless Defense report found only 36% of Australian organizations apply consistent controls to AI agents and nonhuman identities. Cisco's commercial interest is explicit: Solder argues trust must be enforced at the network and security layers, and the company is betting its networking, security and Splunk portfolio becomes the trust layer for agentic AI.