My work domain (an EDU) recently had Google Inbox enabled so I had a good chance to try it out. My personal email is relatively quiet and, I believe, doesn't provide a good Inbox experience. Work is more active and requires actual management, something I've tossed many a tool at over the years. As part of my work life, I supported the Google Apps for EDU installation here and took a pretty extensive presentation to campus about how to manage large amounts of email.
Inbox is a classic Google product: the distillation of a number of excellent ideas into a set of half-complete features built for a use case most people don't meet. We've seen this in the past in products like ChromeVox, Google's Chrome extension for accessibility. ChromeVox works great on ChromeOS devices, but completely ignores the point that most users of accessibility tech (AT) don't have or want ChromeOS devices and come to services with their AT in tow. ChromeVox also ignores decades of convention in AT in favor of simplicity in interacting with Google's own products, allowing Google a quick exit from uncomfortable conversations. ("Google Sheets don't work great with Firefox and JAWS." "It does in Chrome and ChromeVox, use that instead.")
Inbox is also a classic Google product in that it purposefully ignores the successes of a previous iteration (*cough*Hangouts*cough) in favor of a new approach. This is a great strategy for innovation, but leads to frustrating user experiences, confusing messaging about product direction, and destroys confidence that products are built for purpose. Especially when dealing with something as common as email, the use case has to be rock solid before people will try something new. By way of example, let's talk about labels.
Labels were The Innovative Thing when GMail was released (that and storage capacity). They're an amazing management tool which lets folks manage their inboxes in myriad ways. They're flexible enough to allow other products to be built on top of GMail without ruining standard access methods. They provide context, metadata, and the ability to build complex workflows within your mailbox.
In Inbox, labels are far less useful. You can move a thread to a label, but you can't simply apply a label. Labels get used to create Bundles in Inbox, but lose the flexibility found in the standard GMail app to apply visible metadata to conversations. Inbox also does not show you if a label requires attention. For example, if you set up a filter in GMail (which you cannot do in Inbox) to route mail to a label, skipping the inbox, you are never informed in Inbox that the label now contains unread mail. There is also no acknowledgement that people coming from the standard GMail app may have leveraged its features extensively by providing a clear path from one tool to the other. Inbox requires that you manage your mail differently, but does not tell you how to undo your GMail workflows to slot into an Inbox workflow. I suspect Labels exist in Inbox because of someone's impassioned plea to keep them in, despite almost no attempt being made to make them functional.
Inbox also sorts mail into sections loosely based on time (Today, Yesterday, This Month, etc), and then groups Bundles into those sections. But, this model is kind of cheat as bundles within time-based groupings affect the inbox display. For example, anytime a conversation bundled into Low Priority is updated, the entire section moves to the Today group. However, if you expand the bundle, all of the messages are not from Today and the Bundle is sub-grouped by time again. If you clear out the message from today, however, the entire Bundle again moves back down to a time-based section of the last updated conversation. The effect is an entire block of email jumping around in your inbox.
Add to that all of the useful features in GMail lost in Inbox and you're left with a lot of great ideas that almost, but not quite, make a great client. Why, for instance, get rid of Priority markers? Or any of the Labs? The ability to mark a message as unread? Change the theme in any way?
There are answers to all of these, but they all point back to requiring someone rip-and-replace their email workflow completely. GMail has never required that you do that with an app, and it makes it difficult to a) meaningfully test Inbox without destroying your GMail workflows and b) cleanly and easily commit to Inbox.
From experience, we know that Google will iterate on Inbox (although it's been slow to do so so far), but from a team that had a decade or more working with and designing email management tools, the tool is baffling. It is, by its own admittance, incomplete: a link to the standard GMail interface is a default, unhideable link in the top navigation items, allowing you to toggle back and forth between two tools managing your single mailbox. Things you do in Inbox are sometimes reflected in GMail (the default bundles in Inbox and the tabbed labels in GMail), but not always (Pins mean nothing to GMail).
So, where does that leave things? In typical Google fashion, it leaves you with the issue of deciding how to manage your experience. If you want things like Snoozed messages, reminders, and bundled messages, but don't need automation beyond default options, Inbox is great (Snooze is a wonderful feature; I hope they port it to GMail). If you need a labelling system, importance markers, and filtering, GMail is your app.
If you want a mashup of those features, however, you're stuck in the classic Google trap of an innovative-yet-incomplete tool versus a richer experience that misses out on new, useful features.