Even if you're telling the truth, when your security is compromised, so are all your users' email, because you have full access to it.
Bottom line: Yet another "web 2.x", proprietary, "beta," SaaS, subscription, third-party, unrestricted-privacy-compromising tool without a business model that might be around for a few years before being bought and ruined or going bankrupt. One would have to be nuts to trust it and to build one's business processes around it.
And why? Seriously, it could be done client-side without compromising account credentials or email privacy in a bit of Perl or Python, maybe a browser extension, and throw in some Bayesian analysis to reduce the need for manual templating and scraping. A few hours in Emacs/gnus/etc and no need to hand over ALL OF MY EMAIL to yet another "but we're only storing your patterns!" third-party web site.
Someday people will wake up to the insanity and stop trusting these ephemeral convenience services. The question is, will it happen naturally, or will it take more massive security/privacy breaches before people stop being so lazy and nai...I mean, trusting?
People said the same thing when dropbox was announced, look where we are now: very few people do it locally, and dropbox has won.
How is that relevant to the issue of giving a third-party access to ALL OF YOUR EMAIL for the sake of convenience? How does that justify doing so rather than developing a client-side, FOSS solution?
It turns out that Dropbox has set the standard for how to store files on the cloud and sync them for the masses. My gut tells me they're losing grounds to the GDrive/OneDrive/iCloud et al but it's still a default option for many people. In that sense, they have "won" this domain.
Now, don't mistake me: Dropbox and the likes have created a huge problem with data privacy. But as FOSS developers, we must not think that somehow Dropbox doesn't exist or doesn't solve a problem; if we want a better, FOSS alternative, we must set the bar at least as high, and not blind ourselves.