From Hourly to Task Pricing

This is the core monetization problem for AI enabled services, better software makes the work cheaper to produce, but a time based price turns that gain into lower revenue instead of higher margin. In Double’s case, clients buy assistant hours and assistants are paid for hours worked, so every task that AI helps complete faster shrinks both the bill and the worker’s pay. That pushes the whole business toward pricing the finished task, not the labor minutes behind it.

• Double is not selling pure software. It is selling a managed assistant workflow, where a client delegates scheduling, email, research, follow ups, and coordination, and a human assistant uses internal tools to finish the job. That makes hourly pricing feel natural at first, but it breaks as soon as AI handles more of the writing, sorting, and drafting work.
  1 sacra 2 withdouble
• The company was already moving toward fixed prices per task, with examples like asking whether a client would pay $30, $40, or $50 for a completed task, while paying the assistant a fixed amount too. The hard part is not billing logic, it is scoping. Everyone has to agree upfront on what done means before the work starts.
  1 sacra
• This is the same broad shift hitting AI products across software and services. When automation reduces labor time, seat based and hour based pricing start to look backward. Industry analyses increasingly frame outcome based pricing as the cleaner fit for AI, but they also note the operational challenge of measuring success consistently for each task.
  3 ey 4 thomsonreuters

The next phase is likely a hybrid model, fixed task prices for common repeatable jobs, with humans reserved for exceptions and judgment heavy work. If Double can standardize task definitions tightly enough, AI stops being a margin leak and becomes the engine that lets each assistant handle more clients, more consistently, at a lower effective cost.