I launched a wallet last year: https://www.linkedin.com/feed/update/urn:li:activity:7003216362433040384/

Our biggest challenge was by far regulatory. Investors wanted us to be US incorporated, since that's the standard. But US money transmitter licenses are impossible. Our most realistic option would have been using someone like Prime Trust, such that we could piggy back on their licenses. Starting costs for those types of services where $20,000/month. And even at such high prices, you'd still get rugged/shut down by them.

Best approach in hindsight would be waiting for stable-coin type inventions on lightning, such that you can offer fiat balance, but only spendable via lightning rails, and your own responsibility is just reliably running a node. Maybe you can hook into bitrefill type services or disposable digital VISA cards, such that new users immediately know they have somewhere to spend their sats.

Then again, lightning adoption is mad. I'm from a seaside retirement town, and already have a barber, taxi, pizzaguy, gym coach, and cafe all paid for via lightning. Maybe we really don't need fiat rails to be a valid product, especially when Revolut etc add lightning rails.

Although it would be good to add support for a few providers..

Check out this brew package which lets you zap other developers based on the dependencies you're using when developing