CompanyWhat We’re BuildingResearchFoundationStoriesNow at eFindCareersExplore eFind ↗
Products · Stories from Inside

Email has had some time.

Invented in 1971. Still remarkably good at hiding the one attachment you need.

Invented in 1971. Still remarkably good at hiding the one attachment you need.

Email survived because it is useful, open and deeply embedded in work and life.

It also accumulated newsletters, receipts, confirmations, threads, attachments and enough subject lines beginning with Re: to test anyone’s optimism.

It is also where much of modern life quietly leaves receipts: the reservation, the invoice, the shipping notice, the legal attachment, the school message, the project thread and the password reset you needed exactly once.

That makes an inbox unusually valuable as information—and unusually unpleasant as a place to manually excavate information.

We do not think the answer is to make email unrecognizable.

The answer may be to make the information inside it easier to find and understand.

People already understand what a message, reply, folder and attachment are. Reinventing those merely to demonstrate creativity would be a peculiar use of everyone’s afternoon.

The better opportunity is to add intelligence around the familiar structure: find the thing, explain the thread, surface what needs attention and help with the next response.

That is the idea behind eFind Mail.

Ask the inbox. Find the attachment. Understand a long thread. Draft a response. Connect a document to eDrive when the person wants it there.

Ask eFind can become a layer for natural-language retrieval and understanding, while the original message remains available underneath the answer. That matters because summaries are useful; source material is authoritative.

Over time, Mail and eDrive can work more naturally together when the person chooses that relationship. An attachment should not become a permanent orphan simply because it arrived through email.

The intelligence should feel like less work, not a new layer of complexity.

And if the AI needs permission to use something, it should ask rather than develop a creative interpretation of consent.

That includes knowing when not to intervene. A two-word reply does not need a literary editor. A sensitive message may not need a suggested tone. Good assistance is partly the ability to leave the keyboard alone.

We are interested in an inbox that helps people get out of the inbox.

Email has had some time.

We think it is allowed to get better.

More stories