I have been reflecting on a new theme that is emerging. Software was a product. Then it evolved into a service. Now it is a worker.
This is not a re-branding. It is a different category, and it changes what software actually means to an organisation.
Four shifts that come with the new category.
Software does the work, not just records it. The previous generation of enterprise software was a system of record: a place to log what humans had already done. Software-as-a-worker completes the work itself. The output is the deliverable, not the database entry.
It sits inside the org chart, not next to the worker. Software used to live on the desk beside the underwriter, the analyst, the credit officer. Now it sits inside the team, on the rota, in the role description. The credit team has a headcount line for the agent.
The buyer changes. SaaS was sold to the CIO. Software-as-a-worker is sold to the COO and the functional heads. The budget moves from IT to operations, and the procurement template that worked for the last decade no longer applies.
Pricing moves from seats to outputs. Per-seat licensing was a way to charge for access. Workers are charged for the work they produce: memos completed, applications underwritten, exceptions handled. The unit economics of the category change with the unit of pricing.
The skills question.
The interesting question after all of this is what skills software-as-a-worker actually deploys. A horizontal AI worker arrives on day one with no domain context. It can write a polite email. It cannot pass a credit committee.
A vertical AI worker, built for a specific industry and a specific workflow, arrives knowing the credit memo, the covenant call, the KYC file. That is the difference between hiring a generalist temp and hiring a seasoned operator on day one.
What follows the framing.
Treat software as a worker, and the rest of the operating model follows: where you put the budget, who signs the contract, how you measure the output, how you onboard, how you review.
