By Jillian Viner / 13 May 2026 / Topics: Artificial Intelligence (AI) , Generative AI , Cybersecurity , Operational efficiency
Enterprise AI factory deployment has a well-documented failure pattern: organizations procure infrastructure, skip data preparation and governance, take on too many use cases at once, and produce bad outputs from bad inputs. GTT, the third-largest tier-one internet operator in the world, deliberately broke that pattern. After procuring AI factory infrastructure — an NVIDIA and Dell GPU stack deployed across Prague, Slovenia, and New York — GTT took a six-month pause before installing anything. That pause, focused entirely on data governance and use case prioritization, is the foundation of every result the company has achieved since.
James Karimi holds the combined role of CIO and CISO at GTT, a structural decision the company made intentionally to eliminate competing priorities between IT and security. In this conversation, Karimi walks through the full arc of GTT's AI factory journey: the governance model they built before deployment, how they mapped every data source by access method and quality, and the intake process that limits active AI projects to ten at a time — with FP&A review and senior leadership sign-off required before any use case moves forward.
The results are specific and measurable. A finance close process that required 12 to 15 people working across two weeks now completes in one hour using an AI operator. Recruiters who previously spent 60% of their time reviewing CVs now spend that time interviewing — and sales hires are onboarding 8 to 12 weeks sooner, generating monthly recurring revenue earlier. On the network side, GTT is moving toward reducing incident response SLAs from four to six hours down to six minutes using real-time AI automation.
What GTT got right that most organizations miss is the discipline to stop before building. Karimi is direct about the failure rate he observed across peer companies and conversations with NVIDIA and Dell: organizations that skip data governance and take on too many use cases simultaneously are the ones that fail. GTT's 44 acquisitions over eight years gave its leadership team a hard-won appreciation for what happens when discipline breaks down — and that experience transferred directly into how they approached AI factory deployment.
CIOs, CISOs, and IT leaders evaluating AI factory infrastructure will find specific, replicable frameworks in this conversation: how to structure a data matrix, how to run a use case governance model, how to manage organizational change when employees fear displacement, and how to build an AI engineering team from within rather than hiring from outside. GTT's approach is not theoretical — it is already running across three sites and producing results.
If you liked this episode, share it with a colleague.
Have a topic you’d like us to discuss or question you want answered? Drop us a line at jillian.viner@insight.com
James Karimi
CIO & CISO, GTT
Audio transcript:
A full transcript of this conversation will be available shortly.
Marketing Manager, Insight
As marketing manager for the Insight brand campaign, Jillian is a versatile content creator and brand champion at her core. Developing both the strategy and the messaging, Jillian leans on 10 years of marketing experience to build brand awareness and affinity, and to position Insight as a true thought leader in the industry.
CIO & CISO, GTT
Subscribe to our podcast today to get automatic notifications for new episodes. You can find Insight On on Amazon Music, Apple Podcasts, Spotify and YouTube.
