If you worked in enterprise IT during the early 2010s, the current AI discussions feel familiar. Back then, cloud was the strategy every CFO wanted to discuss. Today, AI occupies the same seat.
In many organizations, technology adoption follows a similar arc. A capability becomes widely visible, executives attach efficiency expectations to it, and teams are asked to align. The cloud era produced plenty of strategy meetings and vendor pitches. Some migrations delivered value. Many created complexity that took years to unwind. AI is now entering the same phase.
The risk isn’t the technology itself. It’s the assumption that AI can compress labor costs without a corresponding investment in process design, data governance, and integration. When finance, operations, and customer workflows are mapped independently, automation tends to break at scale. The same was true of cloud migrations: moving an on-premise workload to a provider rarely fixed the underlying process issues.
Executives who see AI as a headcount lever often overlook who will maintain, correct, and govern these systems. LLMs and automation tools still require integration with ERP and CRM environments, data quality controls, and human review for exceptions. Duplicate records, inconsistent approval logic, and fragmented customer lifecycle visibility don’t resolve themselves because a model was introduced. The work changes shape; it doesn’t disappear.
For systems professionals, operations leaders, and technical teams, the rational response is not resistance. Resistance reads as a roadblock. The better path is to understand how the technology works at a functional level—where it fails, how it integrates, and what it can realistically replace. That knowledge becomes leverage, because someone has to keep these systems operational when the narrative cools and the maintenance work begins.
AI is not a workforce strategy. It’s an operational capability that needs architecture, governance, and people who understand the underlying systems. The organizations that treat it that way will get more durable results than the ones chasing the narrative.
