Fetching from the wire…
Top 5 · 2026-05-01 · source-backed
SaaStr published production data from running 20+ AI agents that should make every SaaS founder rethink their product category. Their Salesforce bill went up 80%, from roughly $16K to $22K per year, despite cutting human seats by 60-70%. Meanwhile, Notion usage dropped to literally zero daily active users. Nobody canceled. They just stopped opening it.
The pattern is clean and it's showing up everywhere. Agents use CRM 100x more than humans ever did, hammering APIs and running data operations around the clock. But agents have zero reason to open a collaboration UI designed for humans. Notion, project boards, wikis built for eyeballs. They become invisible.
This isn't just SaaStr's experience. Salesforce's Agentforce hit $800M ARR with 2.4 billion tasks completed in Q4 alone. They're now billing in "Agentic Work Units" instead of seats. Vanta crossed $300M ARR (up 69% YoY) because compliance infrastructure, the stuff that tracks and governs, thrives when agents multiply data access and risk surface.
Oliver Wyman put out a framework that PE firms are already using to triage portfolios: three core SaaS valuation assumptions have broken. Software isn't hard to build anymore. Seat expansion doesn't drive revenue. UI familiarity doesn't create moats. They recommend sorting holdings into "Resilient," "Reinforce," and "Structural Disruption Risk" tiers. That last tier is a polite way of saying "sell before the market figures it out."
The emerging rule is simple. If your product is a system of record, a data store, a governance layer, agents make you more valuable. If your product exists primarily as a human collaboration interface, you're facing what SaaStr calls "stealth churn." Nobody cancels. The usage just evaporates.
I'm building solo, so I feel this acutely. My agents hit databases and APIs constantly. They never open a wiki. If you're building SaaS, ask yourself: does an agent need my product to do its job, or does my product only work when a human is staring at it? That answer determines your next five years.
Each link below shares sources, entities, or timing with this story.
SaaStr published its own invoice, and it's the best single data point I've seen on where agent economics actually lands. A year ago: 10-plus human Salesforce seats. Now: 2 human seats plus 1 API seat. And the bill went up 83%. Not down. Up. The mechanism is that 20-plus AI age...
ServiceNow, Salesforce, and HubSpot have each independently created new revenue metrics that measure AI agent output rather than human user counts. ServiceNow's "Agentic ACV" is at $1B run rate. Salesforce's "Agentforce ARR" hit $800M processing 2.4 billion agentic work units....
IGV is down 24% in Q1. That's the worst quarter for software since 2008. But here's the number that stopped me cold: for the first time in modern history, software valuations have fallen below the S&P 500 multiple. SaaStr put the market cap destruction at roughly $2 trillion s...
Fortune reported this week that CIOs and CTOs are reallocating roughly 40% of IT budgets from traditional SaaS subscriptions to agentic platforms and LLM tokens. I've seen plenty of "SaaS is dead" takes. This one comes with numbers, and the market reacted accordingly. On April...
Per-seat pricing is dying. Not slowly. Not in theory. Right now, across three unrelated categories, simultaneously. HubSpot: $0.50 per resolved conversation. Salesforce Agentforce: $800M ARR on "Agentic Work Units," converting 20 trillion tokens into 2.4 billion discrete measu...
For two years, "agents will compress SaaS seats" was a slide in a venture deck. This week it became a line item in renewal conversations. Enterprises are now openly telling vendors they want to buy 100 seats where they used to buy 500, and vendors are quietly cutting discounts...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.