I suppose it is similar, but you’ll likely need more people per client to make it work. In my experience SaaS software is much easier to scale than SaaS ML. (See also: https://a16z.com/2020/02/16/the-new-business-of-ai-and-how-i... )

On the ML side you probably have more people working on data gathering/cleaning/client specific tweaks than actual modeling. (Practically, I’d say is 80%-20% at best in the projects we do at work).