“Agentic” has become a convenient label for software that can take actions, not just generate answers. The distinction matters. An assistant may draft a customer response; an agent may retrieve the order, issue a replacement, update the record, and notify the customer.
That is a meaningful shift in how work gets done. It is not, by itself, a new business model.
Start with the process, not the promise
The tempting story is that companies will soon run themselves: agents will coordinate, exceptions will disappear, and headcount will become optional. A more useful view is less theatrical. Agents can reduce the cost of particular tasks, shorten handoffs, and make some services available for longer hours. They can also introduce new failure modes: wrong actions taken quickly, unclear accountability, and brittle workflows that break when inputs change.
The practical unit of adoption is therefore not “the company.” It is a bounded process with a clear owner, a defined outcome, and a way to catch mistakes. Start with work that is frequent and rules-governed, but not so consequential that an error becomes difficult to reverse. Keep a person responsible for decisions that require judgment or carry material risk.
The model is not the moat
This framing also changes the investment question. The presence of agents is not a moat. If a tool is broadly available, competitors can adopt it too. Advantage is more likely to come from the less glamorous complements: reliable data, integration into existing systems, well-designed escalation paths, and the discipline to measure whether a process actually improved.
There is a further constraint: customers do not buy “autonomy” as an abstraction. They buy a result, such as a resolved issue or a completed transaction. If an agent makes the path to that result less dependable, lower labor cost is not enough.
The agentic business, then, is not a business without people. It is a business that deliberately assigns work between people and software—and remains accountable for the outcome. That may prove transformative. It is also a more demanding standard than a compelling demo.