On last week’s 20VC pod, @jasonlk articulated a simple product-level insight that can provide the clearest lens I've heard for thinking about the future of all these public SaaS companies. Every one of them will ship some sort of agent feature and that feature will fall into one of two buckets. 1. If the AI agent is “100% good” to quote Jason; i.e. good enough to charge for independently, revenue will re-accelerate and stocks go up 2. If the AI agent is “60% good” and bundled in with the existing product, then the best these companies can hope for is slow revenue growth with declining gross margins. That makes them IBM. And I don't even know the price of IBM. Why would I? What I like about this test is that it cuts through all the SBC/Terminal Value noise and lets you think like a venture investor. Is the agent product good enough that someone will actually pay for it? If a customer is already paying you for “tracking the workflow,” you can only get paid additional money for “doing the work.” Agent revenue is the canary in the coal mine for the future trajectory of all these companies.