Your SaaS portfolio looked fine six months ago. Predictable revenue, predictable churn, predictable earnings calls where CEOs pronounced "AI-powered" eleven times and analysts nodded along. Comfortable.
Then Wall Street stopped believing the script.
By March 31, the iShares software ETF had plunged 21% year-to-date — nearly $2 trillion in market cap, gone. Atlassian, Adobe, Workday: all deep in the red. FinancialContent dubbed it the "SaaSpocalypse." Dramatic. Less so when you open your brokerage app.
The catalyst wasn't one product launch — it was a pattern. Between March 17 and March 31, every major productivity platform pivoted from shipping AI features to deploying autonomous agents that act inside your workspace without asking permission. We've already covered the specific launches and the per-seat pricing death spiral on this channel. The stock market's response, though, tells a different story: when bots do the work, per-seat pricing stops making sense. Wall Street priced that in before most SaaS CEOs finished their keynotes.
The data behind the rout is blunt. A Databricks survey published on March 14 had already shown the trend accelerating: multi-agent adoption spiked 327% in four months, with 78% of enterprises now running at least two LLM families (the AI models powering agents like ChatGPT and Claude). Mistral CEO Arthur Mensch said AI could replace "over 50% of enterprise software." That's not hyperbole from a blogger — that's the CEO of a major AI company calling half your vendor stack dead weight. Bain's March 24 analysis phrased it gently: "Not every software company will be able to make that transition." Translation: some of your vendors are going to die, and they know it.
Here's where it lands for anyone who builds, buys, or depends on SaaS. CIOs are reallocating 40% of IT budgets from traditional SaaS to agentic platforms. Companies using outcome-based pricing — paying for results, not seats — show 31% higher retention, per ainvest's April 4 report. The pricing model itself is mutating from "how many humans touch this" to "what did the software accomplish." For a mid-market SaaS company doing $20–80M ARR (annual recurring revenue), adapting means rearchitecting permissions so agents can act on behalf of employees, redesigning billing for 3 AM bot activity, managing long-running background processes, and absorbing legal liability for what agents do with customer data. That's not a sprint. That's a platform rewrite while the building is on fire.
If you watched this movie between 2011 and 2013, you know the ending. Companies that treated mobile as a feature — slapping a responsive skin on their desktop site — watched mobile-native apps eat their lunch, their dinner, and their market cap. The ones who said "we'll get to it" got to the obituary page instead. The SaaS-to-agentic transition runs the same script, except faster, because agents don't politely wait for your product roadmap to catch up.
Your audit is simple: does your SaaS vendor ship autonomous agents that take actions in your workspace? Not copilots. Not summarizers. Agents. If the answer is "coming soon," start shopping. "Coming soon" is what BlackBerry said about touchscreens.
The SaaS market forked. One half rebuilds around agents. The other half writes blog posts about planning to. Wall Street already named that second group — "legacy." And legacy software doesn't get renewed. It gets replaced.
