Most enterprise AI factory deployments fail before they produce a single result — not because of the technology, but because of what companies skip before they build. James Karimi, CIO and CISO at GTT, shares the governance-first playbook that helped GTT stand up three AI factories across the US, EU, and UK in two months. The results? A two-week finance close reduced to one hour. Recruiters reclaimed 60% of their time in a given week. And that was just the beginning.

Karimi breaks down the deliberate six-month pause GTT took after procuring infrastructure — before installing anything — and why that decision is the reason their AI operators are delivering real results today. He also gets into the combined CIO-CISO role, how GTT handled employee fear around AI job displacement with radical transparency, and what edge GPU deployment will look like in 2026 for enterprise network services.

If you're evaluating AI factory infrastructure, trying to prioritize use cases, or figuring out how to get your organization to actually adopt AI — this conversation has the specifics.

Take the IT Infrastructure Readiness Assessment to see where your organization stands before you build: https://www.insight.com/en_US/what-we-do/expertise/modern-infrastructure/it-infrastructure-readiness-assessment.html

Explore Insight's modern infrastructure expertise: https://www.insight.com/en_US/what-we-do/expertise/modern-infrastructure.html

Subscribe to Insight On for new episodes every week.

Jump to…

00:00 — Introduction and the combined CIO-CISO role
02:09 — What GTT does and its global network scale
02:58 — How generative AI disrupted GTT's business
03:48 — What an AI factory actually is
05:44 — Early wins from three AI factory deployments
07:24 — The six-month data governance pause
09:58 — How GTT prioritizes AI use cases