Skip to content
Episode 742

Build an AI Revenue Management Agent in 30 Minutes

November 16, 2026 Jasper Ribbers
Listen Now

Get Your Free Revenue Report

Want to optimize your STR pricing? Get a personalized revenue report from our experts.

Get Started

Want to outperform the market? Freewyld Foundry’s Revenue and Pricing Management service is driving an average 18% performance lift for STR operators, even in down markets. If you are managing 15+ listings and want a free pricing audit, apply here

Jasper Ribbers replays his live webinar on building an AI revenue management agent from scratch, using nothing but a prompt and two exports from PriceLabs, no code or technical experience required.

You will learn:

  • Why revenue management was never about finding the “perfect price,” and how the chess blunder, mistake, inaccuracy framework helps you prioritize which prices to fix first
  • The 26 factors Jasper’s own team weighs on every pricing decision, uncovered when an intern’s master’s thesis studied their own revenue strategy
  • Exactly what to export from PriceLabs, neighborhood data, the multi-calendar, and booking history, to build a working AI pricing agent
  • How that data becomes an agent that flags your most underpriced and overpriced dates, then builds custom reports like a holiday pacing report in seconds
  • Why AI should never run unsupervised, and how Jasper still corrects his own agent’s recommendations after eight months of daily use

We also talk about:

  • How to run this without PriceLabs, using the exported data from any pricing tool
  • Whether AI is replacing pricing tools outright, or just the person who used to stare at the calendar
  • How to tell a pricing problem apart from an Airbnb visibility problem

Mentioned in the Episode:

The STR Pacing Playbook cover
Free Guide

The STR Pacing Playbook

A step-by-step guide to tracking and using pacing data to optimize your short-term rental pricing strategy.

Download Free Guide

Introduction

Jasper: If you're not using AI in your revenue management, you're missing out, because the power of these tools is pretty incredible. What I love about something like Claude is you don't really need to learn how to use it. You can just ask it how to connect to your pricing tool and it'll tell you. Today I'm going to show you how to build your first revenue management agent in about 30 minutes. We built a prompt that gives you the basics of what to start with, and from there you can build on it. It's going to analyze every open date on your calendar for the next six months, or twelve if you want, and tell you where to focus your attention. You don't need to be technical. You don't need any code.

A little about me: I'm Jasper, founder and head of revenue strategy at Freewyld Foundry, and host of the Get Paid for Your Pad podcast. We've been doing revenue management for about three years now, closing in on 90 clients and a quarter billion dollars in managed revenue. Everything I'm sharing today comes out of that experience.

Why There's No Perfect Price

Jasper: First, what is revenue management? There are a lot of definitions, but it's not really about finding the perfect price, because the perfect price doesn't exist. Every unit is unique, every day is unique, and you only get to sell a given night once. You can't rewind time and test a different number.

If you ask our ten revenue managers what the best price is for one night on one unit, you'll get ten different answers, maybe $190, maybe $210. But none of them will say $50, and none of them will say $500. So the real basics of revenue management are making sure every price on your calendar is defensible: high enough that a booking at that price is a win, low enough that someone looking in your area would still consider you.

We've analyzed portfolios doing over a million dollars a year, and in pretty much every one, we find obvious mistakes. That's where the chess analogy comes in. Learning chess isn't about finding the perfect move, it's about eliminating blunders. Even grandmasters have inaccuracies. If you just avoid the blunders, you're already in the top 10%. It's the same with pricing: get reasonably priced across your whole calendar, and you're already ahead of most operators.

The Blunder, Mistake, Inaccuracy Framework

Jasper: We make educated guesses on pricing. We look at comparable listings, how the unit booked last year, demand pacing, how far out we are in the booking window. One of our interns, doing his master's thesis on short-term rental revenue management, came up with a list of 26 factors we actually weigh when we price a unit. I was kind of blown away by that, 26 factors, and we never really know if we're right. That's the irony of revenue management: it's a lot like trading.

Say a similar unit is priced $140 to $280, last year this date booked at $200, we're 80% full one month out, and it's early, 120 days out. What's a reasonable price? Not $150, that's clearly a loss. Somewhere between $210 and $280 is where most people would land, even though nobody agrees on the exact number. That's where AI comes in: we can teach it what a reasonable range looks like, then have it scan your entire calendar every day and flag anything outside that range.

If you have 25 listings with 12 months open, that's over 9,000 individual price points. No human is reviewing that by hand, and it all changes daily: new bookings, market shifts, check-in rules. Most operators don't even have someone dedicated to pricing. It's whatever the general manager or co-founder can squeeze in once the fires are out, usually just the next two weeks, which isn't necessarily where the biggest opportunity actually is.

Building the Agent

Jasper: To build this, all you need is Claude or ChatGPT. I use Claude. You need the prompt, which we shared live during the webinar, and you need to know where to pull your data from.

Start a brand new project in Claude with no other context. If you're on PriceLabs, go to the neighborhood data section and download it for each listing, along with the multi-calendar export and your full booking history, booked, not cancelled. Upload all of that into the chat along with the prompt, and Claude analyzes your entire calendar and tells you where it thinks you're priced too high or too low.

Its first pass won't be perfect, it doesn't know your specific strategy yet, so you review its recommendations, check the actual prices, and give it feedback. For example, it flagged me as underpriced on a date that was only two days out. I told it that for underpriced properties within the last 14 days, it's fine to drop to our minimum. That's the kind of feedback loop you build over time. I still use the blunder, mistake, inaccuracy labels to prioritize what to look at first.

Reports On Demand

Jasper: Once Claude has your data, you can ask it for all kinds of reports. I asked it to build a Thanksgiving pacing report with our own branding, nights sold, revenue, ADR, pace versus last year, and it had it ready in under a minute. You can use the same thing for owner questions: why is this unit priced at $400, how did it do last year, just hand that to Claude and let it answer.

I also had it build a simpler version: current revenue on the books for November, December, and January versus the same time last year, plus how much unbooked potential is left and what percentage you need to fill to hit last year's number. That's a genuinely useful way to think about pacing: once you know how much unbooked revenue you have at the start of a month, you start to get a feel for what's realistic to book by the end of it.

Audience Q&A

Kaye: We got a great question from Tom in the Netherlands: how are you actually using this day to day with client portfolios?

Jasper: Our internal tool turns every data point from your pricing tool into different views based on feedback from our revenue managers. Our daily workflow starts with a quick health check on the portfolio, pacing for this month, next month, two months out, and flags any listings falling behind on pickup over the last week or 30 days. It also checks Airbnb listing visibility, since losing bookings isn't always a pricing problem. Then it surfaces its top recommended price changes, which we can push live through the API, plus new bookings and cancellations, and anything else it thinks we should know about since the day before.

Kaye: Lisa asked, what if you're not a PriceLabs user, and is AI basically replacing pricing tools altogether?

Jasper: Every pricing tool has an API and the same underlying data, PriceLabs just gives you the most control, so you're not locked out if you use something else. And no, AI isn't replacing the tool. It's a recommendation layer on top of it. Think of the pricing tool as the race car, you're the driver, and the AI is the coach helping you drive better. You need all three.

Kaye: Heather's asking whether this agent runs in the background on its own, or if you have to prompt it, plus how to set it up in Claude.

Jasper: You keep working inside one ongoing project or thread so Claude keeps the context of everything it's learned about your portfolio. Each day you feed it new data, your latest bookings, and if you're connected through an API or MCP it can pull that automatically. Context management matters a lot here: you can attach documents to the project with specifics about your portfolio as you notice where it's making mistakes. And yes, you can ask Claude to set up a recurring task so it pulls data or generates a report on a schedule without you prompting it every time.

Kaye: And Matthew asked, with everything changing in the Airbnb algorithm, how do you tell if a booking problem is about price or about distribution?

Jasper: If your price is competitive and you're still not getting booked, it's usually a visibility issue, not a pricing one. Sometimes a listing becomes invisible on Airbnb, you search in a private window and it's just not there, and you have to call Airbnb support, I've seen that happen more than once. The way we built our own tool, it flags when a listing stops getting picked up, checks whether the pricing explains it, and if not, looks at the listing itself for things to fix. Bad reviews matter here too, a run of one or two star reviews will tank your visibility just as much as a pricing problem will.

Closing

Jasper: Definitely get into AI if you haven't already. It saves an enormous amount of time, and you'll fall behind if you don't start now. Thanks so much for showing up today, it was great to see everyone.