The Last General-Purpose App

Before we bury Microsoft 365 and Google Workspace, let's be honest about the numbers: Microsoft still owns 58% of the global productivity market. That's 446 million paid seats. Google holds another 28-ish percent, dominant among startups and remote-first companies. Together, they control roughly 86% of how the world creates documents, sends emails, and builds spreadsheets.

That's not a market ripe for overnight disruption. That's a market with extraordinary inertia.

But here's what the market share numbers obscure: both Microsoft Copilot and Google Gemini are creating walled gardens that make it progressively harder for vertical apps to integrate deeply. The choice between Microsoft and Google in 2026 is no longer about features — it's about which AI ecosystem you want locked into. And that's the tell. When the selling point shifts from "what it does" to "what it prevents you from leaving," you're looking at a platform in defensive posture.

The incumbents will survive. The question is whether they'll survive as productivity tools or as AI distribution platforms that happen to include a word processor.

Indigo, a platform built specifically for medical professional liability workflows, just closed a $50 million Series B — oversubscribed. Not a general "AI document tool." Not a "smart spreadsheet." A purpose-built system that understands malpractice laws, liability data structures, and the specific cadence of how medical legal teams actually work.

The round was oversubscribed because investors see what enterprise buyers are discovering: a tool that knows your domain eliminates an entire category of human translation work. You don't need to explain to Indigo what a certificate of merit is. You don't need to build a custom template in Word and hope your paralegal fills it in correctly. The domain knowledge is the product.

This is the pattern playing out across legal, healthcare, real estate, and manufacturing: the moment an AI tool can speak your profession's language natively, the general-purpose alternative starts feeling like a foreign-language dictionary — technically useful, but exhausting to work through. And when the stakes are medical malpractice, "good enough" isn't.

Perplexity just made its AI browser Comet free on Windows, macOS, and Android. That's notable on its own — giving away a fully agentic browser is a land-grab move. But the real story is what Comet does: it replaces tabs with tasks. The browser doesn't just display your email, calendar, and documents. It manages them. Autonomously.

Perplexity calls this "Agentic Browsing," and the name is exactly right. Comet doesn't wait for you to click. It reads your inbox, drafts replies, schedules meetings based on context, and manages your task queue — all within the browser itself. The browser becomes the productivity suite, and the productivity suite becomes... optional.

This is perhaps the most radical threat to general productivity software: not a better spreadsheet or a smarter doc editor, but a layer that sits above all of them and coordinates the work that previously required you to context-switch between six different apps. If the browser is the new OS, then Office and Workspace are just filesystem endpoints.

Uptool raised $6 million in seed funding to build what it calls a "vertical operating system" for small and mid-sized manufacturers. The thesis is bracingly simple: the factory floor runs on paper clipboards, email chains, and Excel files not because those tools are good, but because nothing better existed. Asana and Monday.com were built for knowledge workers. They don't understand lot tracking, quality holds, or production line dependencies.

Uptool's bet is on "compound workflows" — the multi-step, multi-person processes specific to manufacturing that break down when forced into generic project management software. An order comes in, triggers a material check, which triggers a production schedule, which triggers quality checkpoints, which triggers shipping. In Excel, that's five separate spreadsheets maintained by five different people who email each other. In Uptool, it's one flow.

This is where the "end of general productivity software" thesis gets most interesting: not in Silicon Valley offices where people already use 47 SaaS tools, but in industries where the default was always "paper plus Excel" because horizontal software never bothered to learn the domain. Vertical AI isn't replacing existing software there — it's replacing the absence of software. And that market is enormous.

On February 3rd and 5th, the software sector experienced what traders are calling the "SaaSpocalypse." Salesforce, ServiceNow, Atlassian, and their peers lost a combined $285 billion in market capitalization. The trigger? Growing investor conviction that agentic AI — particularly tools like Anthropic's "Claude Cowork" — will fundamentally break the per-seat subscription model that underpins the entire SaaS industry.

The logic is straightforward and devastating: if an AI agent can do the work of three account executives in Salesforce, you don't need three Salesforce licenses. If a code agent can manage a project in Jira, you don't need the project manager's seat. The per-seat model assumed that software productivity scaled linearly with humans. AI agents break that assumption.

Whether the sell-off was overdone (it probably was) matters less than what it signals: the market now believes that general-purpose SaaS is vulnerable to AI substitution. That belief becomes self-fulfilling. It changes how investors fund, how enterprises budget, and how startups position.

Microsoft confirmed what everyone suspected: commercial prices for Microsoft 365 are going up 5-33% in July 2026, depending on the plan tier. The justification? "Expanded availability of AI features." In other words, you're paying for Copilot whether you asked for it or not.

The immediate effect is a forced "value audit" across every enterprise IT department in the world. When your Microsoft bill jumps by a third, someone in finance is going to ask: "Do all 5,000 employees actually need the full Office suite?" And that question — which nobody bothered asking at $13/seat/month — suddenly has a very different answer at $18.60.

Here's the irony: Microsoft is bundling AI into everything to prevent customers from leaving for standalone AI tools. But by raising prices to fund that bundling, they're giving every CIO a concrete dollar amount to justify exploring alternatives. The sales team doesn't need Copilot in Word — they need a purpose-built AI that knows their pipeline. The legal team doesn't need Copilot in Excel — they need a tool that understands contract risk. For the first time, the cost of general productivity software is creating the business case for vertical alternatives.

Microsoft is inadvertently subsidizing its own disruption. And the July price hike hasn't even hit yet.