Now that I think about it, I am not sure why they didn't just announce using the smartphone speaker - I've seen several vendors with smartphones still using these boxes.

Granted, many vendors don't have smart phones, and perhaps sound boxes started with them. I think the fintech companies saw the opportunity and decided to not implement it their app for users with smartphone. Given that they are so cheap, I guess the vendors even with smart phones decided to get them anyway.

It's difficult to wake up a phone to play custom audio. I believe iOS and Android has made that impossible without rooting the device.

You don't need to play custom audio. Notification narration already exists.

Yes, the vendors have smartphones. The smartphones' speakers are not loud enough to announce over the road noise and general buzz in Indian shops. It would also be too distracting to keep looking down on the phone.