On Oct 27, Greg Isenberg (@gregisenberg, about 752K followers), founder and CEO of Late Checkout, posted on X: "saas is becoming free: give away the crm, give away the dashboard, give away the app, and charge for the agent that does the work." Within about seven hours the post had about 109K views, 1,526 likes, 277 replies and 37 quotes, with many founders debating whether per-seat pricing can really give way to pricing for the work an agent completes. This is a discussion of an opinion, not a pricing announcement from any company.

Key Takeaways

  • ✓Engagement (as of 2026-10-28 03:10 UTC, public X API data): about 109K views, 1,526 likes, 746 bookmarks, 277 replies, 37 quotes
  • ✓The claim: the CRM, dashboard and app become free, and the paid product is the AI agent that does the work for the user
  • ✓The most common pushback: charging for "work" means defining in writing what done means, who checks it, and a monthly cap, or it cannot be billed
  • ✓Another camp says companies will keep paying for the system of record, while SaaS vendors give away MCP access so outside agents can read and write
  • ✓This is a founder opinion under discussion, not a pricing change by any company
🧭

Turn your technical choice into a development budget

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

🔬

In-Depth Technical Analysis

Background and pain points

This is an X discussion, not a pricing announcement. On Oct 27, Late Checkout CEO Greg Isenberg posted a few lines: SaaS is becoming free, so give away the CRM, the dashboard and the app, and charge for the agent that does the work. It touches a worry SaaS companies have had for two years: AI has made building a CRM far cheaper, and per-seat pricing is getting harder to defend.

Architecture highlights

The post splits a product into two layers. The interface and data layer are free and act as the way in. The paid part is the agent on top that finishes tasks for the user, such as following up with customers, updating the CRM or producing reports. Quote posts compared it to a free phone with a paid plan, or a cheap printer with expensive ink. Others argued the opposite: companies really pay for the system of record, and SaaS vendors will give away MCP access so agents from OpenAI or Anthropic can read and write.

Benchmarks

There are no benchmarks or public revenue numbers here, and Isenberg did not name any company doing this; one reply asked who does it well and has no answer yet. As of 03:10 UTC the post had about 109K views, 1,526 likes and 277 replies.

Getting started for developers

The most common concerns in the replies are measurability and trust. Charging for work means writing down the task scope, who checks it and a monthly cap, and every agent action needs to be logged, traceable and reversible, especially where mistakes are expensive, like payments or payroll. If you build agent products, start recording how many tasks of each kind get completed and how often a human has to review them; that data is what outcome-based pricing will need.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

Up NextX discussion: Dario Amodei says some Claude tokens are worth cents and others tens of millions, floating pay-for-results AI pricing
Read Now→