For a service like this, I can simply use the 'concierge MVP' (https://www.shortform.com/blog/concierge-mvp/). So you're not far of: Rather than an army, it's me myself and I. Another common strategy is to do things that don't scale (and this most definitely won't scale). If it turns out to be a success I can define with the customers which features will be implemented first, and which will remain a 'concierge' service. The technical risk will need to be tackled next when there is market demand.
For early adopters, that pretty much means they're getting a personal coach for near to nothing.
If I'm lucky to have customers, I have some ideas around converting chat bots into virtual coaches, doing a bit of natural language processing (mainly to get the sentiment) and some simpler text matching to make sure the target customer, problem and solution match. Under the hood I'll need a process with different branches for different businesses. The alpha stage is also meant to further expand this process before implementing it in software.
(sidenote: Launching something is exciting, but also scary. I'm super happy you wrote!)