Notion Kills Its Email App as Users Switch to AI Agents
Notion is shutting down its Skiff-influenced email client, citing a shift in how users manage their inboxes — most are now routing through AI agent workflows instead of the app directly. The move signals Notion's bet that agentic inbox management, not another email client, is where the product should go.
Original sourceNotion is pulling the plug on its email application, an effort that drew heavily from Skiff's design DNA after Notion acquired the privacy-focused productivity startup in 2024. The company says the core reason is simple: the majority of active users who connected their inboxes were doing so to feed AI agents, not to use the email interface itself. When the surface layer goes largely unused because users have already routed around it, the honest move is to kill it.
The shutdown reflects a broader pattern in productivity software right now. Point-and-click email interfaces made sense when the job was reading and filing messages manually. But as AI agents get better at triaging, drafting, and acting on email autonomously, the traditional client becomes an unnecessary layer. Notion is essentially saying it doesn't need to be in the email UI business — it needs to be in the email data and automation business, where agents do the lifting.
For users who relied on the email app as a unified inbox inside Notion, this is a disruption. Notion is reportedly directing them toward its agent-based tooling and API integrations, though the transition path varies depending on how deeply embedded their workflows were. The Skiff acquisition was pitched partly as a privacy-forward expansion into communications; sunsetting the email app less than two years in suggests the thesis shifted faster than expected.
The decision is notable not just for what Notion is killing, but for what it reveals about the trajectory of inbox products generally. If a company with Notion's user base and resources can't justify maintaining an email UI because agents are doing the job instead, that's a signal worth paying attention to — both for competing email startups and for teams still building traditional inbox experiences.
Panel Takes
The Skeptic
Reality Check
“Notion bought Skiff, stripped its privacy DNA for parts, shipped an email client nobody asked for, and is now blaming AI agents for the fact that users didn't adopt it. 'Most users use AI agents instead' is a graceful way to say the product didn't win on its own merits. The question worth asking: are users actually running sophisticated agent workflows, or did they just never open the email tab?”
The Founder
Business & Market
“The Skiff acquisition cost real money, the email app cost engineering cycles, and the moat they were building — a privacy-first unified workspace with communications — just got abandoned eighteen months in. Pivoting to 'agents handle your inbox' is a fine thesis, but it's also a product that needs to monetize differently than a stickier email client would have. The expand story from document collaboration into agent-managed communications is plausible, but Notion needs to ship agent tooling that's actually defensible before this bet pays off.”
The Futurist
Big Picture
“The falsifiable thesis here is: by 2027, the dominant email interface is not a UI but an agent loop — read, prioritize, draft, send, with humans approving rather than driving. Notion killing the email app isn't a retreat, it's an early data point confirming that thesis is already happening at scale among their user base. The second-order effect nobody's talking about: if agents become the primary email client, the companies that own the context graph those agents run on — documents, tasks, meeting notes — hold structural leverage over the companies that own only the inbox.”
The PM
Product Strategy
“The job-to-be-done for the email app was 'manage your inbox without leaving Notion,' but if users were connecting email primarily to feed agents, the actual job was 'give my AI the right context to act on my behalf.' Those are two completely different products with two completely different success metrics, and building the first one when users wanted the second is a classic job-to-be-done miss. Notion should name the specific agent workflows that are replacing the email UI before it kills the client — otherwise this is just a shutdown with a good PR frame.”