Software pricing is shifting from seats to work performed. On the latest edition of Syft Data’s GTM Field Notes, Patrick Thompson explains the model behind Clarify. Give away the CRM software, then charge based on the work the AI actually does. Unlike traditional SaaS, every AI action carries a real inference cost. That makes usage-based pricing less of an experiment and more of an economic necessity. The future of software may be free tools, paid labor.

Build versus buy is not a new debate created by AI. On the latest edition of Syft Data’s GTM Field Notes, Patrick Thompson and Imran Patel discuss how companies were asking the same question at Snap nearly a decade ago. AI has changed how quickly and cheaply companies can build software. The fundamental decision has never changed.

On the latest edition of Syft Data’s GTM Field Notes, Patrick Thompson explained why MCP is powerful, but still far from reliable infrastructure. MCP tool calls are expensive, brittle, and prone to breaking changes. Unlike traditional API integrations, they often lack the contracts and stability companies expect from production software. Agents may look magical in a demo. Keeping them working in the real world is a very different problem.

On the latest edition of Syft Data’s GTM Field Notes, Patrick Thompson made the case for having strong opinions about your product. He shipped, then unshipped, three features in a single week because they failed to meet the quality bar. Shipping fast matters. Knowing when something no longer deserves to exist in your product matters even more.

On the latest edition of Syft Data’s GTM Field Notes, Patrick Thompson made a great point about the economics of AI. AI is cheap right now because venture capital is subsidizing our addiction. But the addiction is real. Once you experience the productivity gains, going back to working without AI feels impossible. The real question is what happens when the subsidies disappear, and prices begin to reflect the actual cost of delivering all that intelligence. A lot of AI companies are effectively spending $2 to make $1. The winners will be the companies that preserve the productivity advantage without inheriting an unsustainable cost structure.

Danny Chepenko explains why bigger prompts don't automatically produce better copy. In fact, they often make it worse.

Your CRM shouldn't just track deals. Danny Chepenko shares why it should be the source of truth for every customer interaction. Think of it like GitHub, but for sales.

Every founder obsesses over CRM hygiene. Sometimes it matters less than you think. When demand is overwhelming, even messy systems can keep up. Danny Chepenko on why product demand beats perfect operations.

Most AI companies are selling tokens. Customers don't care about tokens. They care about outcomes. Danny Chepenko explains why the future belongs to companies that sell value, not usage.

Most founders think cold outbound is either high volume or highly personalized. It's not. The best teams know when each approach makes sense. Here's how Danny Chepenko thinks about high-volume vs. low-volume outbound.