Part 2 of 2. Read Part 1 here: Why SaaS Is Breaking
Every disrupted company says this. Every single one.
Chegg said it about homework help. "Our expert network provides nuanced educational support that AI can't match." Then ChatGPT launched. Revenue collapsed.
SaaS companies are saying the same thing right now. "Our workflows are too complex." "Our compliance requirements are too specific." "Our integrations are too deep." I've heard those lines in investor calls, board presentations, and due diligence reports. Always the same structure. Our thing is too hard for AI.
The blind spot is universal. People who built something complex can't imagine it becoming simple. They confuse the difficulty of creating something the first time, when the patterns didn't exist, with the difficulty of replicating it once agents can reuse those patterns at machine speed. A project management product took years to build from scratch. Now it takes minutes to vibe-code a credible substitute. The complexity didn't vanish. The cost of handling it did.
This is where investor behavior has to change. Once you accept that the old moat logic is unstable, the question stops being "is AI important?" and becomes "what do we do differently with capital, diligence, and operating plans right now?"
This isn't just a threat story. It's an opportunity story too, and most investors are missing that half.
The average mid-market company spends $4,000 to $10,000 per employee per year on SaaS subscriptions. A 500-person company burns $2-5 million annually on project management tools, CRM, analytics platforms, content software, helpdesk systems, collaboration suites.
Agents replace entire categories of that spend.
I run six specialized agents every day. They handle work that used to require dedicated SaaS products: content operations, financial analysis, research, project management. I didn't cancel those subscriptions because I found a cheaper vendor. I cancelled them because the products stopped being necessary.
That makes the math compelling for portfolio companies. Redirect 30-50% of SaaS spend into agentic infrastructure and you can self-fund the transition. No special budget. No heroic transformation program. Just cancel the tools your agents already replaced and move the money.
For companies buying SaaS, this is the biggest cost-optimization opportunity in a decade. For investors holding SaaS companies selling into that spend, it's a revenue risk most financial models still understate.
The SaaSpocalypse hasn't hit every category yet. Some sectors are already broken: simple project management, basic CRM, commodity analytics, content generation tools. Those businesses are done or dying. The market has already repriced them.
But complex vertical SaaS, deep workflow automation, regulated sectors, and companies with genuine data-network effects still have 12-18 months before agents fully catch up. Not forever. But enough.
Enough to map exposure honestly. Which portfolio companies have moats that hold? Which ones are still running on assumptions agents have already invalidated? Be ruthless. "Our product is too complex" is not a moat. "Our product operates in a regulatory context where hallucinations could kill patients, and we built ten years of compliance infrastructure" might be one.
Enough to restructure pricing. Per-seat pricing is dead for any product agents can interact with directly. If your customer's agent workforce scales from 5 to 500 without buying another seat, your pricing model breaks before your board deck catches up.
Enough to pivot positioning. Companies that move from "tool you use" to "infrastructure agents depend on" still have a path. Companies that keep selling features to human users are selling into a shrinking market.
The companies that assess agentic exposure now, honestly, with someone who actually builds, still have a survival path. The companies that wait until multiples are already compressing are Kodak, except they don't get Kodak's 15-year head start.
The gap isn't information. Everyone can read about AI disruption. The Wall Street Journal covers it. Gartner publishes frameworks. McKinsey releases decks. There is no shortage of commentary.
The gap is evaluation capability.
When a portfolio company CEO tells you, "our product is too complex for AI to replicate," how do you test that claim? You can't. Not unless someone on your team actually builds with autonomous agents and can pressure-test that product in an afternoon. Someone who can look at the architecture, the workflows, the value proposition, and tell you plainly: this moat holds, or this moat is paper-thin.
That person doesn't exist on most investment committees. The partners understand markets, capital structure, and operations. The operating advisors understand go-to-market, pricing, sales efficiency, org design. Nobody in the room actually builds with the technology that's doing the disrupting.
The cost of adding that capability is trivial compared to one quarter of value erosion in one portfolio company. The ROI is obvious. The only question is whether you hire that capability before the write-down or after it.
Two trillion dollars gone. Revenue multiples cut in half. Enterprise seat declines showing up in earnings. A $5 billion main product functionality replicated in days.
The structural shift is no longer theoretical. The window is open, but it's closing.
Assess your portfolio's agentic exposure with a practitioner who builds, not a consultant deck. Map which moats hold and which are paper-thin. Stop telling your board you're "exploring AI."
Start acting.
The SaaSpocalypse is real. It is structural. And it is still early enough to respond. Not for much longer.
The investor lens I've used throughout this article has a natural endpoint: the portfolio company that actually builds the agentic-native operating model. Not the one that adds an AI feature. The one that redesigns around agents from the ground up.
What does that company look like?
- fewer SaaS subscriptions, replaced by agent infrastructure
- engineering focused on platform, security and governance, not individual application delivery
- business units operating with direct agent access, no intermediary ticket queues
- operational cost structures that no longer scale linearly with headcount
The investment question for the next cycle isn't which SaaS companies survive. It's which portfolio companies build the agentic-native model fast enough to protect their margins and extend their moats.
For any company to move toward that natural endpoint, the first step is not a strategy offsite. It's a trustworthy due diligence assessment of the company's exposure to agentic working, so the transformation actions that follow are grounded in reality rather than slogans.
I cover what that model looks like in practice at Agentic Native. For practitioners and investors who want specifics, not frameworks.