The Skiff Saga: A $0 Acquisition That Ended Twice

Notion acquired Skiff in February 2024, absorbing the encrypted productivity startup’s engineering team and technology into its own operation. The price was effectively zero — an acqui-hire framed as a strategic talent grab rather than a product purchase. Within months, Notion killed Skiff’s native email service, wiping out @skiff.com addresses and leaving users who had built their communication around the platform scrambling for alternatives.

That should have been the end of the story. It wasn’t.

The ex-Skiff engineers who joined Notion didn’t move on to unrelated projects. They built Notion Mail, a Gmail client that launched in April 2025 and carried the clear DNA of their previous work. For the small but dedicated user base paying attention, Notion Mail was Skiff’s second life — same team, same institutional knowledge, different logo. Productivity tool enthusiasts who had migrated away from Skiff watched the new email client as a kind of resurrection.

Notion shut it down anyway. The company announced in mid-2025 that Notion Mail would go dark across web, desktop, and iOS on September 22, citing low adoption among its core user base. The shutdown notice claimed most Notion users don’t rely on traditional email clients at all, handling electronic correspondence through AI agents instead.

Most coverage treated the Notion Mail closure as a standalone product failure. That framing misses the actual story. From acquisition to first shutdown to relaunch to second shutdown, this is one continuous 18-month arc — a single team’s work acquired, repurposed inside a larger platform, and ultimately abandoned twice. The @skiff.com addresses are already gone. The Notion Mail inbox goes dark in September. Users who trusted either product at any point in that cycle lost their investment in the workflow twice over.

The pattern matters because it reveals how acqui-hires actually function inside large productivity software companies: the talent gets absorbed, the product becomes a feature experiment, and users become test subjects with no guaranteed continuity.

The AI Agent Excuse: Legitimate Pivot or Convenient Cover?

Notion’s official explanation for killing Notion Mail centers on a single claim: most of its users have already moved to AI agents to manage their email, making a dedicated inbox app unnecessary. That framing deserves a hard look.

AI agent adoption for email management is, by any honest measure, still in its earliest stages. The users autonomously delegating inbox triage to AI tools today represent a narrow slice of the productivity software market — early adopters, developers, and technical power users. The idea that this behavior describes “most” users of a general-purpose productivity platform like Notion strains credibility. Notion has never published adoption metrics, usage data, or any figures that substantiate the claim. No percentage of active Notion Mail users was cited. No benchmark for what “most” actually means was offered.

What the framing accomplishes, strategically, is significant. Instead of acknowledging that Notion Mail failed to carve out meaningful share in a Gmail client market already dominated by Superhuman, Spark, and Google’s own interface, the shutdown announcement positions Notion as a company that read the future correctly and acted accordingly. The product didn’t lose — it simply became obsolete before the competition could finish it off. That is a more comfortable story to tell investors, users, and the press.

The backstory makes the spin harder to ignore. Notion acquired Skiff in February 2024, shut down Skiff’s own email service within a year — eliminating all @skiff.com addresses in the process — and then launched Notion Mail in April 2025, built largely by the engineers who came over through that acquisition. The entire arc from acquisition to shutdown spans roughly 16 months. Launching a product and discontinuing it that quickly rarely reflects a calculated response to shifting user behavior. It reflects a product that did not gain traction.

Notion may well be right that AI-driven email workflows will eventually make standalone inbox clients redundant. That prediction could age well. But deploying it as the primary rationale for a shutdown, without data and less than a year after launch, reads less like strategic foresight and more like reputation management.

What This Means for the Gmail Client Graveyard

Notion Mail joins a graveyard that keeps filling up. Google killed its own Inbox app in 2019 after users built entire workflows around it. Astro shut down in 2018, barely a year after launch. Newton cycled through shutdowns and revivals before going dark again. Superhuman survived by charging $30 a month and targeting a narrow slice of power users who would actually pay — a model almost nobody else has replicated. The pattern is consistent enough to be a structural verdict: third-party Gmail clients face a market that systematically destroys them.

What makes the Notion Mail shutdown different is the exit narrative. Previous email app deaths came with straightforward admissions — not enough users, not enough revenue, too expensive to maintain. Notion’s announcement took a different route, claiming that most of its users now rely on AI agents to handle email rather than traditional inbox clients. That framing is doing heavy lifting. It lets Notion exit a failed product while positioning the failure as forward-thinking rather than market rejection. Users who spent months building Notion Mail workflows cannot easily push back against a company that says the entire category is obsolete.

That accountability gap matters beyond Notion specifically. When any productivity tool can wrap a shutdown in AI disruption language, the feedback loop that punishes bad product decisions breaks down. Investors hear a pivot story. Users are left holding broken workflows with no meaningful recourse.

The Skiff acquisition makes the economics starker. Notion bought Skiff in February 2024, absorbed its engineering team — engineers who had built a well-regarded encrypted productivity suite — and then shut down Skiff’s own email service, eliminating @skiff.com addresses in the process. The Notion Mail app those engineers built lasted roughly five months before the September 22 shutdown announcement. That timeline, from acquisition to product death, signals that even experienced teams with real technical credibility cannot manufacture sustainable demand for standalone email clients in a market owned by Gmail and increasingly crowded by AI-native alternatives. Resources and talent were not the missing ingredient. The market itself rejected the category.

The Real Cost: Users Who Trusted the Product

Some users absorbed this shutdown twice. When Notion acquired Skiff in February 2024, Skiff’s existing users lost their @skiff.com email addresses as the service wound down. Many of those displaced users migrated to Notion Mail when it launched in April 2025 — a reasonable bet, given that Notion Mail was built largely by the same engineers who came over through the Skiff acquisition. Now those users are being asked to migrate again, with Notion Mail shutting down across web, desktop, and iOS on September 22, 2025.

That is not a single platform failure. That is the same failure, twice, to the same people, inside an 18-month window.

Most coverage of Notion Mail’s shutdown focuses on the product decision itself — the claim that Notion users now prefer AI agents to handle email rather than a dedicated client. That framing treats the shutdown as a rational pivot. It ignores the people who built real workflows around the tool, who configured their inboxes, set up automations, and trusted that a product released just months ago would have a meaningful lifespan.

Platform dependency carries a hidden cost that rarely appears in shutdown announcements. Users invest time learning a tool, migrate their data into it, and reorganize their habits around its specific interface. When the platform disappears, that investment doesn’t transfer. The friction of rebuilding a productivity workflow isn’t trivial, and for users who went through it once with Skiff and again with Notion Mail, the damage compounds.

The broader signal here matters for anyone currently relying on Notion’s ecosystem. The acquire-rebrand-shut down pattern Notion has now executed on email gives users of Notion Calendar, Notion AI, and the core workspace product a concrete reason to ask whether their workflows are safe to depend on long term. That question has no reassuring answer yet. Notion has not addressed the trust deficit this pattern creates — only the product rationale for the latest shutdown.

Notion’s Larger AI Bet and What It’s Actually Building Toward

Shutting down Notion Mail frees the engineering talent Notion absorbed through its February 2024 Skiff acquisition — a team that built encrypted email and productivity tools before Notion folded them in. That team is almost certainly getting redeployed, not disbanded. The September 22 shutdown date isn’t a retreat; it’s a resource reallocation disguised as a product decision.

The strategic logic Notion is betting on runs something like this: the future of productivity software isn’t a better email client or a smarter document editor — it’s an AI orchestration layer that sits above all of those apps simultaneously. Rather than competing with Gmail on inbox management or with Microsoft Word on document creation, Notion wants to be the AI agent that coordinates across email, calendar, docs, and databases on a user’s behalf. That’s a fundamentally different product ambition than building best-in-class tools for individual categories.

Notion’s own announcement made this explicit, claiming that most of its users already rely on AI agents to handle email rather than traditional clients. Whether or not that statistic reflects the broader market or just Notion’s heaviest power users, the company is treating it as a directional signal and building accordingly.

The harder question is whether Notion can actually compete at the AI agent layer. Google has Gmail data, Calendar data, and decades of communication graph intelligence baked into Gemini. Microsoft has Copilot wired directly into Outlook and Teams. OpenAI is building operator-style agents with access to far more compute and training data than any productivity startup can match. Notion’s advantage — a flexible, deeply interconnected workspace where users already store projects, notes, wikis, and databases — is real but narrow.

The pivot trades a losing battle in email clients for a harder war in AI-powered productivity automation. Notion is making the bet anyway. Whether that’s visionary repositioning or a company chasing a trend it can’t outrun depends entirely on execution over the next 18 months.

The Broader Signal: AI Is Becoming the Reason Products Die, Not Just Live

Notion Mail’s shutdown on September 22, 2025 marks something more significant than one failed product. It represents a turning point in how AI gets used as a business narrative — not to justify building, but to justify quitting.

For years, “AI-powered” was the phrase that unlocked funding, user interest, and press coverage. Companies launched products because AI made them possible. Notion’s explanation for killing its own email client flips that logic entirely. The company told users that most of them rely on AI agents to handle email correspondence anyway, making a dedicated inbox redundant. AI didn’t kill Notion Mail through competition. Notion used AI as the reason to walk away from its own investment.

That sets a template other productivity software companies will copy. Watch for a wave of shutdowns between now and the end of 2026, each one framed around some version of “AI agents have made this product category obsolete.” Task managers, note-taking tools, scheduling apps, lightweight CRMs — any product sitting in a category that AI agents can theoretically absorb becomes a candidate for this kind of exit. The announcement language will sound forward-thinking. The underlying reality will be a company cutting costs and narrowing focus while using AI as the cover story.

This changes the due diligence every user and business needs to run before committing to a productivity platform. The old question — does this tool use AI? — is now nearly useless as an evaluation criterion. The sharper question is whether the company building this tool sees AI as a reason to abandon the product category it currently occupies. Notion acquired Skiff in February 2024, launched Notion Mail in April 2025, and shut it down within months. That is a complete cycle — acquisition, launch, shutdown — compressed into roughly 16 months.

Anyone building workflows inside productivity ecosystems needs to treat AI agent roadmap language as a risk signal, not a feature benefit. When a company starts describing its own existing tools as friction that AI will eliminate, users are looking at a product that’s already been internally written off.