What is Stopping Most AI Success?
What is stopping most AI success?
McKinsey’s rule of thumb for agentic AI is $1 on the tech, $3 on the process, and $5 on getting people ready. Most companies I meet spend it the other way around.
The technology matters, and companies pay Workfast to teach them to build, build with them or build for them. I am not arguing to slow the build. What I see, over and over when arriving at a new client or prospect, is a stack of AI and agents and little redesign of the work those agents are supposed to change. Roles, processes decision rights, and the org chart stay the same. Then leadership asks why utilization is flat and funds another pilot.
McKinsey’s August piece on the agentic adoption gap is blunt about this. Successful transformations follow that 1:3:5 pattern. Most companies invert it and treat capability building as an implementation detail. In their August State of Organizations research, leaders named change management and siloed ways of working as bigger obstacles to scaling AI than the technology itself. Microsoft’s notes point the same way: value shows up when the way of working changes, not when another tool lands in the catalog.
In my experience at
Workfast Consulting, the gap is rarely I picked the wrong tech. It is unused software. Teams build agents, watch them sit idle, and respond by building more. That is an output problem dressed up as progress. Outcomes live somewhere else: which use cases move the needle, what a role looks like once an agent takes the grunt work, and what in the operating model has to change so people trust the handoff.
Workfast sits in a unique spot with both tech and transformation. Not many people keep pace with the technology and also know how to reimagine work, roles, and structure. That is how we help clients. The goal is not a shinier AI or a cooler tech stack. It is a new way of working to drive results in this golden age of AI, which only happens if organizations spend on the people side of the ratio.
If your team is sitting on AI licenses or agents nobody uses, I am glad to compare notes, and glad to talk with people in your network who want to accelerate the value, not the volume. I will put the McKinsey link in the comments.
Do you see organizations focusing their money and focus on the wrong part of the problem? Do you agree with the $1/3/5 model? What do you see as the barrier to not realizing AI results?
Learn more about James Stovall, the co-founder of GA AI Alliance here and here!
You can book a conversation with him here.
