Hey all, We recently launch our new pricing for our CPQ platform. I'd love to get everyone's perspective on it. I'd welcome any candid feedback. You wont hurt my feelings. (much) https://www.getveles.com/pricing

Saw a few people here asking about Claude skills for pricing, which got me thinking. Triggered by a discussion in this post and my own project, I've created a small tool. Nothing fancy – just a structured way to think through pricing and packaging without overcomplicating it. If you want to play with it: https://github.com/Billing-Atlas/billing-diagnostics-skill Your feedback is welcome and appreciated!🙂

Howdy pricing people! Wrapped a great Office Hours session last week with David Reid and Sam Little from Teneo on AI pricing strategies. David's been doing software monetization for 17 years and said the quiet part out loud — "you've had to relearn everything you thought you'd learned over the last 18 months." That pretty much framed the whole hour. A few things that stuck with me: 1️⃣ The disconnect is real. David's team asks enterprise buyers if they want value-based pricing. They say yes, absolutely. Then Teneo comes back with a consumption model and the same buyers say "no chance, never getting this past the CFO." He pointed at Twilio and Snowflake as examples of companies that had to rebalance toward hybrid after going too hard on consumption. The "buyers say one thing, buy another" gap is just louder than ever with AI. 2️⃣ The gentle on-ramp he keeps using with $1B companies Embed a meaningful chunk of AI credits into the base license, and design the allowance so most customers don't exceed it in year one. Only the real power users blow through it. Removes friction, drives adoption, buys you a year of usage data to set the real bands on. This seems like an increasingly popular approach. Clay's recent remodel and introduction of "actions" comes to mind. 3️⃣ You earn the right to monetize AI by proving value first. David's been running a pattern with clients: three-month free trial, translate credits into value during the trial, then start charging in month four. And the business case doesn't have to come from new AI revenue — one billion-dollar client of his saw GRR go 5% higher on AI users vs. non-AI users. So $50M of their "AI revenue line" came from retention lift, not new ARR. I hadn't thought about AI ROI that way before. 4️⃣ The hub-and-spoke GTM model. For big portfolios launching a bunch of AI features at once, Sam's push is a small central "control tower" (product, pricing, product marketing, sales, revops) that sets the value metric, pricing architecture, and commercial guardrails — then the product teams run as spokes inside those rails. The alternative is either siloed chaos or one giant bureaucratic committee. Both lose.

Wanted to sanity check a view with you: MCP matters product-wise, but not as a standalone pricing unit. My take is SaaS companies shouldn’t monetize MCP itself. It’s a connectivity layer, and customers don’t buy “protocol access” - they buy the value it unlocks. So the monetization likely shows up in: - premium AI layers — paid add-ons / tiers with custom agents, integrations, context, admin controls, or workflow automation - usage-based pricing — credits, actions, outcomes - seat expansion / plan upgrades - pull-through to core product usage So my thesis is: don’t monetize the protocol; monetize the value created through it. Curious what others are seeing: - what usage patterns are real? - how are you packaging it? - what are customers actually paying for?

Does anyone have any insights into AI coach pricing? My product group is working on a product using a 3rd party service and they have quoted us $0.23 per interaction for 965,000 interactions in a year. We are using our own data and I would say the AI coach is pretty standard. My product people think that pricing is too high. I would appreciate any insight from anyone on this.