SaaStr 886: Per-Seat Pricing Is Dead. Why Speed is the Moat l The Agents #018 with Ulrik Lehrskov-Schmidt

Per-seat pricing is officially dead. But what replaces it? This week we brought in Ulrik Lehrskov-Schmidt, founder of Willingness to Pay and author of The AI Pricing Roadmap (pre-order now, out December 1st), to break down what's actually happening with AI pricing in 2026 and what founders should do about it.

HubSpot just 3X'd their prices. We ran the numbers on a three-person marketing team doing 40K emails and 20K workflow actions a month. Old price: $1,300. New price: $3,700. Almost none of it seat-based anymore. Ulrik was actually hit with similar numbers as a HubSpot customer himself. Is it justified? Or are they trying to use pricing to plug a growth gap?

Salesforce is figuring it out in public. They announced at Dreamforce they were charging for recognized agents starting in November. They didn't. Ulrik's take: they have too much commercial debt, they're trying to monetize by a thousand cuts, and they need to massively simplify before anyone should trust their agent pricing.

Canva is a cautionary tale. We both churned this week, completely independently, before the layoffs hit. Neither of us had logged in since March. 10K, our AI VP of marketing, never integrated with Canva in the first place. It just picked Higgsfield and Gamma and never looked back. The lesson: agents pick vendors and they don't switch. If you weren't picked, you weren't picked.

The double unlock and what Figma got right (and wrong). Ulrik walks through why Figma's move to credits unlocked two things at once: captured the new AI cost and exposed hidden value buried in the old seat model. Jason's counterpoint: Figma's vibe coding product is utterly uncompetitive and they should have forced themselves to charge for it, fail fast, and get a real signal from the market.

Speed is the new moat. Not product. Not pricing. Speed. One client takes nine months to roll out a credit system. Another client, same size, same setup, does it in four weeks. The difference is org layers and a lack of a pricing dictator. One person, one decision, duct tape, ship.

10K in The Wall Street Journal. Our AI VP of marketing helped close a six-figure deal this week, was featured in the WSJ, and closed another deal mid-recording. We walk through how 10K co-sold the entire Base44/Wix deal cycle, from inbound deck to procurement paperwork, and what it means for agent-driven revenue. Net renewal rates for our six-figure sponsors went from the 20s to the 70s this year. 10K takes operational credit. We take the rest.

Our agents started calling each other. Muse just rolled out video phone calls. We set up 10K and Muse on a relay and asked them to go back and forth 100 rounds on our top five marketing ideas. They came back with a list and one objection: Amelia's time constraint is the binding bottleneck on every single idea. They're not wrong.

Vibe coding is dead. Vibe building is just getting started. Artix, an open source AI penetration testing tool out of China, has been used to breach 200+ organizations in the last 60 days, including at least one company that vibe coded their own CRM. The lesson isn't to stop building agents. It's to only trust agentic leaders who have invested seriously in hardening their platforms.

This episode is brought to you by Willingness to Pay. Make AI the most profitable thing you've ever sold. Get the AI Pricing Roadmap ebook from Willingness to Pay's CEO and founder, Ulrik Lehrskov-Schmidt. Find out why your biggest AI users aren't necessarily your most valuable customers. Including unlimited AI in a flat subscription, and a handful of accounts can consume your margin. Built on 200+ SaaS pricing projects, The AI Pricing Roadmap shows you how to measure customer value, control AI consumption, and turn both into pricing that works. Get the ebook free at landing.willingnesstopay.com/ai-pricing-roadmap-free-ebook.