what could each individual Fedimint provider offer to potential users to lure them into using their minted tokens

1. The user knows someone from the mint personally. This is central in Obi's view.
2. The mint performed the trusted setup ceremony correctly and convincingly.
3. For users that want the mint to keep their private key, geographical proximity because the user will have to go to half of the minters personally. Unfortunately, those users are going to suffer a lot when (maybe if, but most likely when) the regulators crack down on the mint. If the minters are gathering together at regular intervals to receive users who want to restore their keys, cracking down on the mint becomes much easier.
4. For users willing to manage their private keys themselves, the opposite of geographical proximity. Ideally, one minter in US, one in Europe, one in China, one in Africa etc. Put other OpSec stuff in this category.
5. Healthy cap policy. If there's no cap, there might occur a situation when the mint holds $1T and hundred criminal organizations are trying to rob it. If there's a cap, there might occur a situation when a user wants to receive $10 but can't because the mint doesn't want to mint anymore. Ideally there should be a cap but the amount of minted tokens should not be close to it.
6. Stablecoin integration. Taro is mentioned in the article but it's barely started; RGB is already released although nobody understands it. But in principle there could be already stablecoins on RGB and lightning-integrated minted stable tokens. There is also USD-denominated BTC-backed tokens from Kollider but I'm not quite sure about those.
7. Custom tokens. "The mint has pleasure to announce that at Bob's request we minted 100 BobCoins redeemable from Bob for 1 hour of his work per coin". This is very important for poor people who want to borrow but can't accept the risk of currency volatility (and I don't mean BTC volatility only). Obviously more important for small mints.
8. Other bells and whistles. AMMs between BTC and stables. Lending. NFTs because why not. Some analogue of ln.cash would be very nice for tipping.
9. LN providers: the amount and quality of LN providers servicing the mint. It's best if their identity is not known since the mint doesn't have to trust its LN providers; but then it might occur that several LN providers are in fact the same person.

And how would a profit-seeking Fedimint group charge a fee for their service? You provide them a 100 sats and get the equivalent of 99 sats back in their own mint?

1. the exchange of Bitcoin for tokens and vice versa, but other people could compete with them on that
2. if they run contracts for people in the mint, they could charge a fee for that I suppose too