Ask HN: App monetization strategies in the AI era?
I'm building an tool that combines OCR and text categorization of an image. Each job accrues cost with compute time. The rest of the tool is negligible cost.
The outcome is not discrete, so I worry about charging a usage based fee. If the output is not perfect (9/10 times), I wouldn't ask a user to pay for that job. A subscription model is an option, but a user could easily outpace the monthly cost if processing many jobs in a month.
What are some useful monetization strategies you've seen for AI tools to overcome this challenge? I haven't seen it but tokenization economics strategies might be an area to check out. I don't have a good source to reference. Why can't it run locally? It may be easier than you think to do the inference on the phone.