A small number of companies are building entirely new things with AI. Most businesses are doing something different: transforming how they do the work they already do. New technology, same domain.
Two Very Different Stories
When a major technology arrives, there are always two kinds of companies.
AI-native startups, foundation model labs, new tool companies. These organizations are building products that genuinely didn't exist before. New markets, new capabilities, new business models. For them, AI is the core product.
Law firms, accounting practices, marketing agencies, manufacturers, logistics companies, retailers. These organizations aren't changing what they do — they're changing how they do it. The domain stays the same; the execution layer shifts. This is the story for most businesses.
The headlines tend to focus on the first group. But if you work in an established business, your story is almost certainly the second one.
The Outsourcing Parallel
This is not the first time an industry transformation looked like existential disruption but turned out to be something else.
The same product. Made differently, in a different place, at lower cost. The product category didn't disappear — production moved offshore. Companies that adapted found new roles in design, quality, supply chain, and brand. The domain survived; the execution changed.
The same software. Developed by different teams, in different time zones, at different price points. Software development didn't disappear — the work redistributed globally. Engineers who adapted found themselves specializing in architecture, product, and the work that couldn't be standardized.
The same legal analysis, the same financial reports, the same customer support, the same code. Done differently — with AI handling the volume and pattern-matching, while humans provide judgment, relationships, and accountability. The domain doesn't disappear; the execution layer shifts.
In each case, the core insight is the same: the domain didn’t change. The execution layer did.
What Actually Changes
When businesses transform rather than get replaced, the shape of work shifts — not the existence of the work.
- A lawyer still does legal analysis. The AI drafts; the lawyer reviews, refines, and takes accountability.
- A designer still does design. The AI generates variations; the designer applies taste, judgment, and client knowledge.
- A support agent still serves customers. The AI handles volume and drafts responses; the agent handles escalations and relationship nuance.
Some tasks within roles become less prominent. Others — the judgment calls, the relationship work, the accountability — become more central. This is how it has always gone with transformative technology.
ATMs didn’t eliminate bank tellers. When ATMs reduced the cost of running a branch, banks opened more branches — which meant more tellers overall, doing different work: less cash-counting, more relationship banking. The same pattern has appeared across every major technology shift.
The Right Question to Ask
The conversation most organizations are having — “will AI replace our jobs?” — is the wrong starting point. It generates fear and defensiveness, neither of which helps.
The more useful question is: “How do we redesign our work around what AI can now do?”
That’s a question with productive answers. It focuses attention on process redesign, skill development, and where human judgment adds irreplaceable value — rather than on anxiety about outcomes that are, in most cases, not the actual risk.
📝 Key Concepts
- Two tracks: builders creating new things vs. transformers doing existing work differently — most businesses are in the second group
- New technology, same domain: the execution layer changes; the core work and its value persist
- The outsourcing parallel: manufacturing, IT services, and now knowledge work all followed this pattern
- Work changes shape, not existence: tasks shift, judgment becomes more central, accountability stays human
- Ask the productive question: “How do we redesign our work?” moves forward; “will AI replace us?” generates fear