Service as a Software is a market trade, not a margin trade

SaaS was a margin trade. It captured the software dollar — a thin slice of the total spend in any category. Service as a Software targets a fundamentally different prize.

Sequoia's framing is simple: for every dollar a business spends on software, it spends roughly six on the people who deliver the service that software supports. SaaS captured the dollar. Operators captured the six. Two decades of SaaS valuations were built on the one. Service as a Software is the first model that lets a single company credibly target all seven.

The bet is that AI is now good enough, in a narrow but expanding set of workflows, to be the operator itself. Instead of selling an accounting plugin to a bookkeeper, the company sells the finished tax return directly to the small business owner. The output approaches what a human accountant would have shipped. The price drops by close to an order of magnitude.

That conditional matters: in the categories where the workflow can be fully owned by an agent. The full-seven argument is the upper bound, not the central forecast. The realized share will be smaller and slower than the bull case, particularly in regulated categories. But the structural claim holds — the addressable market isn't a slice of labor spend. It's the entire labor spend in a category, repriced.

Source claim: SaaS captured only the software dollar; Service as a Software targets the full 1:6 ratio by replacing the human operator, not just tooling the human operator.