Table of contents
Office Hours
02
Pricing AI Usage: MCPs, Unlimited Plans and the Sticky Questions Underneath
Most of the questions teams bring to AI pricing right now (how to charge for MCP access, whether to keep unlimited plans, how to handle agents and fair use) are really one question wearing different hats. This Office Hours session worked through the lot. The short version: those are all the same problem, and they resolve the moment you decide whether you are selling a door or the work that goes through it.
Date
June 18, 2026
Format
Online, live Q&A
Registered
32
Questions
30+
Overview
- Thestarting point: Why AI usage pricing feels harder this time, and the threemoves that cut through it.
- Your questions, answered: Fourteen questions from the session,grouped into four themes, answered in Ulrik's own framing.
- Your next steps: The six-step sequence to run with your team in thenext 90 days.
- One map to keep: The single decision (door vs work) the whole modelfalls out of.
- Work with us: How Willingness to Pay designs AI pricing end to end.
Half thequestions in this session were one question wearing different hats. MCP,agents, unlimited plans, fair use: they all ask the same thing, which is whatdo I charge for when the thing using my product might be a machine, the costscales with use, and the buyer still wants a bill they can predict. The answerruns on one distinction. An MCP endpoint, an API, your own interface: those aredoors into your product, and you do not price the door, you price what getscarried out through it.
Decide whether you are selling an input (tokens, calls,raw access, priced like a utility) or an outcome (a job done, priced on value).Then wrap whatever you land on in a unit the buyer can actually budget, whichfor most B2B AI is a credit, and put a fair-use limit in the contract beforeyour heaviest users find the hole in it. Cost sets your floor. The value chainsets your metric. The buyer's budget sets your unit.
Get the report
Download the full Office Hours report (PDF)
What is inside:
- The "door vs the work" map for deciding what to meter through MCP,your API and your UI.
- Fourteen real attendee questions answered on MCP metering, agents, credits,unlimited and fair use.
- The 90-day sequence for repricing AI usage without overcomplicating it.
- Ulrik's published frameworks, with public examples only and no customerdata.
If you only remember three things
- Price the work, not the door. MCP, API andUI are channels into the same product. Decide the value metric once and letevery door meter the same thing, or you will rebuild your price list every timeyou ship a new interface.
- Decide input or outcome, then picka unit they can budget. Inputs (tokens, calls) are a utility you price thinand on volume. Outcomes (resolved tickets, processed documents) are priced onvalue. Either way, wrap raw usage in credits or workflows the buyer can measure,predict and control. Nobody budgets tokens.
- Sell unlimited to humans, never totheir agents. An agent does not self-regulate, so "unlimited" withno limit in the terms is a blank cheque. Set a soft fair-use limit where 95 to98% of customers never touch it, and a hard safety stop behind it.
Three lines from the session
"An MCPendpoint is a new door into your product. You do not price the door. You price what people carry out through it."
NAME
"An agent is just a user that never sleeps, never gets bored, and presses the button a thousand times a second. Price the work, and cap the downside."
NAME
The team behind the session
.jpg)
Ulrik Lehrskov-Schmidt
CEO, Founder
A globally recognized authority on pricing strategy and the author of The Pricing Roadmap . With over 20 years of experience, Ulrik has been a trusted advisor to globally recognized SaaS companies navigating complex pricing transformations. He holds a graduate degree in Analytical Philosophy and an MA in Finance from Harvard.
.jpg)
Morten Klank
Senior Pricing Advisor
A seasoned SaaS executive with over 20 years of leadership experience. Morten brings extensive experience leading organizations through complex change, strategic repositioning, and pricing optimization. He has headed several high-impact turnarounds in PE-backed companies, aligning commercial models with long-term business strategy. Morten holds executive education from INSEAD in innovation and change leadership.
Roee Hartuv
Senior Pricing Advisor & Head of GTM Expertize
A B2B SaaS executive and seasoned advisor, Roee Hartuv brings over 20 years of experience driving revenue growth at high-growth companies. He has led sales, customer success, and marketing teams across multiple scaleups and led the consulting practice at Winning by Design. Roee is based in Berlin and holds an MBA from Tel Aviv University and a BA in Computer Science.
Christopher Truce
COO & Co-Founder
Christopher Truce is a seasoned expert in financial services and SaaS technology products, with a proven track record in product development, commercial strategy, and pricing innovation. With over 18cc years of experience across global markets, Christopher excels in aligning pricing models with complex customer product ecosystems and driving adoption in highly regulated Industries. Chris holds an MA in business management and an EMBA from Stanford University.
Need help?
Rethink how much you charge.
We design AI pricing models end-to-end: strategy, packaging, metrics, guardrails and implementation. We specialize in transformation, not optimization. We can't help you tinker with your pricing. But if you're ready for a redesign, connect with us.
200+
Pricing redesigns
242%
Average uplift
125
Avg. days to launch
0
Failures or blow-ups
