Most enterprise agent conversations start in the wrong place. Someone asks for a platform demo, a roadmap, a model comparison. I have sat through enough of those to know none of it tells you whether the organization can put an agent into production. The first useful move is narrower: scope the AI stack that already exists.
That stack is rarely a diagram anyone maintains. It is a coding assistant adopted team by team, a pilot trapped in one business unit, a chatbot wired to a wiki, automations nobody has touched since the builder changed teams. Tools got bought. Experiments got run. Almost none of it was operationalized. The Operationalization Gap: Where AI Demos Go to Die is not a slogan. It is what you find when you sit down and inventory the work.
An assessment is not a courtesy you offer before the real pitch. It is the first piece of work that looks like software. You name the workflows still living in inboxes and standing meetings. You separate demos from systems. You mark where context actually sits, which jobs can run in the background, what the reviewable output should be, and where a human still has to approve before anything counts as done. Specialized work stays specialized. One mega-prompt is not an operating model. If you cannot draw that, you do not have an agent strategy. You have a tool budget.
Spend the hour on their stack, not yours. A point of view up front is enough. The rest is argument: what is already running, what is rotting, what a human must still approve, which workflows are ready to become software. A capability tour does not answer those questions. A map does.
Do the inventory before the roadmap. A thin map means the pilot was never a product. A thick map tells you which agents to build, what context they need, and where approval sits. That is the deployment. Everything else is a demo.
Key takeaways
- Enterprise agent work starts by inventorying the AI stack already in place, not by demoing a platform.
- An assessment is the first unit of real work, because it separates demos from systems and names where context and review sit.
- If you cannot draw the stack, you do not have an agent strategy. You have a tool budget.
FAQ
What belongs in an AI stack assessment?
The tools already adopted, the pilots that never left one team, the workflows still done by hand, where context actually lives, and which outputs a human must approve before they count as done.
When is a company ready to deploy agents?
When it can draw the stack. Readiness is a map of workflows, context, and review, not a successful demo or a model selection.
Related Essays
The Operationalization Gap: Where AI Demos Go to Die
The gap between an AI demo and an AI deployment is called software engineering. Most organizations are not equipped to close it, and that is where all the value lives.
The Agent Lab Replaces the Agent Deploy
Enterprise agents need a living environment where users iterate on behavior, not a one-time deployment pipeline.
Start With the Pain, Not the Platform
The temptation with agent projects is to start with the technology. That is how you end up with a demo that impresses nobody who actually has to use it.