AI is everywhere—until it isn’t. For many organizations, what begins as a promising AI proof of concept fizzles before it reaches production. In fact, as Deloitte’s Steve Brown noted in a recent webinar hosted by Deloitte and VAST Data, over 90 percent of AI POCs never scale.
So what’s standing in the way?
Brown joined VAST's John Mao to peel back the layers of complexity behind the AI Factory, a framework for scaling AI innovation across the enterprise. The conversation dove deep into why so many initiatives stall, including three critical roadblocks:
Lack of specialized talent to build and manage AI agents,
Limited compute availability at scale, and
Data and compute sovereignty requirements that introduce cost and compliance challenges.
But this isn’t a doom-and-gloom story. It’s a blueprint. The panel shared how VAST and Deloitte have partnered to design pre-optimized AI infrastructure environments - tested and benchmarked across real-world workloads - that deliver both performance and cost efficiency. According to Brown, customers should expect to see a 40–50 percent lower TCO and 30-40percent faster performance by deploying an AI Factory model built around VAST’s Insight Engine and Deloitte’s silicon-to-service framework.
What does that look like in practice? The webinar walks through powerful, real-life use cases, from automating pre-operative patient Q&A in hospitals to massively accelerating product development cycles using a network of generative AI agents that integrate across data silos. These aren’t hypothetical deployments; they’re active, scalable solutions driving measurable business impact.
Perhaps most importantly, the session explores how to work with your existing data, wherever it lives - on-prem, in SaaS platforms, across clouds - without initiating a massive (and expensive) migration effort. The hybrid-ready architecture of the AI Factory means that organizations can finally deploy inference, training, and retrieval workloads where the data is, not where they wish it were.
If you're wrestling with how to take AI beyond experiments and into everyday operations, this webinar is a must-watch. Watch it here.
