📌TL;DR — At Scale UK 2026 in Brighton, I sat in on “The Shifting Role of a Revenue Manager” with Oliver Stern (Wheelhouse), Ishan Gangakhedkar (Strategy Cues), and Eduardo Mandri (Sentinel), moderated by John An (TechTape). The verdict from people managing thousands of listings: AI is absorbing the reporting, the monitoring, and increasingly the day-to-day pricing moves — but not the judgment calls. LLMs give a slightly different answer every time you ask, so deterministic models still rule pricing decisions. Owner trust, risk profiles, and local context remain human work. Revenue managers aren’t disappearing — they’re managing far larger portfolios.

I was in the room in Brighton for this one, and the question hanging over the session was blunt: now that AI is everywhere, what’s left for the revenue manager to do?

John An opened with a short history of the profession that rang true to anyone who’s been around since the early days. The Excel power users were once the experts. Then dynamic pricing tools — software that automatically adjusts nightly rates based on demand, seasonality, and market signals — changed the game. Revenue management emerged as a concept around 2019, and by roughly 2024 it had become a formal discipline that most professional operators had built into their businesses.

Technology has already reshaped this role once in a decade. Now AI is doing it again — and as An put it, half-joking: “Maybe it’ll destroy it. I’m not sure.”

Here’s what I took away, and what I think it means for property managers.

1. AI Summarizes the Data — Humans Still Make the Judgment Calls

Everyone on stage agreed on one thing: data gathering is a solved problem. The pickup reports (how many bookings came in over a given period) and on-the-books spreadsheets (reservations already confirmed for future dates) that once defined the job are now automated.

What still separates top operators — in 2019 and today — is interpretation.

Gangakhedkar made the point that judgment calls sometimes mean deciding against what the data appears to say. AI can summarize what’s out there; it can’t tell you when to overrule it. First-principles thinking still applies.

Mandri, whose team has managed revenue for thousands of listings over seven years, drew the sharpest technical line of the session: feed your entire portfolio to a large language model (LLM) — the technology behind tools like Claude or ChatGPT — and ask what’s right and what’s wrong, and it will give you an estimation, and a slightly different answer every time you ask. His team’s split is clear: AI to understand what’s happening in a portfolio, deterministic and machine-learning models to decide what changes. (Deterministic means the same question always returns the same answer — exactly what you want when setting prices.)

Stern’s test was the simplest: when you know what you want to calculate, the answer should be repeatable.

Between the lines: the hype says “let AI set your prices.” The practitioners actually running thousands of listings still draw a hard wall between LLMs for analysis and algorithms for decisions. Consistency is a feature, not a limitation. Worth remembering next time a pitch deck tells you otherwise.

2. Owner Relationships Remain Stubbornly Human

If there was one theme the panel kept circling back to, it was this: owner communication is the part of the job AI can’t own.

Stern put it concretely: even if rate changes become fully agentic — meaning the AI takes the actions itself, not just recommends them — someone still has to understand the data well enough to tell an owner, “We’re 2% down, but the market is 5% down.” That takes clear benchmarking (measuring your performance against the market and comparable properties, not just your own history) and real confidence in your numbers — and it applies internally too, with your sales and operations teams.

Mandri added the dimension I found most interesting: owner risk profiles. Some owners fixate on occupancy — they just want to see the calendar full. Others only care about ADR (Average Daily Rate — the average nightly price guests actually pay). Neither is the optimal metric, but it’s what they want. Algorithms assume everyone wants maximum revenue; the reality is messier. We fall into the trap of thinking every owner is a pure profit-maximizer. They’re not.

And An offered the warning I’d put on a poster: tell an AI to “maximize my revenue” without rules of engagement, and it may slash rates to fill the calendar — technically hitting the goal while destroying exactly what the owner cares about. Your owner calls you up: “What are you doing? That is not what I wanted.”

What I’d do with this:

• Treat each owner’s risk tolerance as a data point in your RMS setup (your revenue management system — the software that sets and adjusts your prices) — segment your portfolio and configure pricing, cancellation, and length-of-stay policies per layer.
• Benchmark portfolio vs. market, not just portfolio vs. last year.
• Never hand an AI a goal without guardrails.

3. Local Context Can’t Be Scraped

Gangakhedkar’s example landed with the whole room: construction work next to one of his properties. How is AI supposed to know that? There’s no feed for it.

A human has to enter that context into the RMS, adjust the listing description, and warn guests before arrival — rather than pray for a decent review at checkout.

His broader point deserves more attention than it got: operators with two to three years of actualized data — real results from stays that actually happened, not forecasts — are sitting on an asset most of them underuse. You can establish your own cause-and-effect patterns — what actually happened when your gap-night policy changed (how you price the orphan nights stranded between two bookings)? When your minimum length-of-stay rules moved from X to Y? What was the impact of those last three poor reviews?

Pattern recognition is getting easier by the month. But as Gangakhedkar stressed, having the information is not enough — it has to become an action point for your revenue team and your operations team.

4. Fewer Spreadsheets, Bigger Portfolios

Nobody on stage predicted the death of the revenue manager. They predicted leverage.

Mandri shared a number I noted down: his team of seven revenue managers has seen productivity rise dramatically over the past 12 months, simply by letting the tools surface the signals that need attention. His conclusion: revenue managers aren’t going away — they’re handling far larger portfolios, and they have to.

Stern offered the most useful mental model of the session — three layers of how AI advances the role:

1. Querying data — prompting for exactly the numbers you need instead of hunting through software menus.
2. Building interfaces — creating the business intelligence (BI) dashboard you want instead of waiting on a vendor roadmap.
3. The agentic layer — AI no longer just answering questions but taking the actions itself, on top of that data and those interfaces.

He also flagged where the role is heading: closer integration with distribution and marketing, especially as direct bookings grow.

Between the lines: there’s a trap inside the efficiency gain. A more scalable revenue function risks becoming a more siloed one — and Stern said as much. The revenue manager of 2026 needs to be more connected to marketing, distribution, and operations, not less, even as AI shrinks the manual workload.

5. Optimize Revenue, Not Occupancy or ADR — and Manage by Lead Time

An audience member asked the question every owner eventually asks: should I optimize for occupancy or ADR? It drew the clearest tactical advice of the session.

Stern’s answer: it all depends on lead time — how far ahead of the stay date you are.

Stern’s lead-time playbook:

• Far out: de-risk the calendar by accepting bookings at slightly less aggressive rates — locking in a revenue base early.
• Compression window: when demand outstrips supply and the market is filling fast, push ADR as high as your booking pace allows (pace = how quickly reservations are coming in versus the same point in previous years).
• Last minute: shift back toward occupancy — an unsold night is revenue you never get back.

Mandri was blunter, and I agree with him: good revenue management always optimizes revenue. Occupancy and ADR targets exist only because some owners demand them. The ideal client is the investor who simply says: maximize the return on this asset — and then lets you do it.

Bonus: How to Hire a Revenue Manager — One Test from Each Panelist

A property manager in the audience asked how to vet revenue manager candidates. Each panelist gave one test, and together they make a decent interview kit:

• Stern: have candidates walk through their actual workflow and explain it in layman’s terms. You’re testing the work and the owner-facing communication skills at the same time.
• Mandri: probe their willingness to learn from the machine. He’s seen experienced revenue managers so stuck in their ways that it took months to accept a counterintuitive recommendation that turned out to be right.
• Gangakhedkar: ask them to articulate a failure — “I missed this target because I didn’t do A, B, C” — and what they learned from it.

My Take

The panel’s answer to “will AI destroy the revenue manager?” was a clear no — but I’d add a nuance. AI is absorbing the monitoring, the reporting, and increasingly the day-to-day pricing moves. What it isn’t absorbing: judgment, owner trust, and local knowledge.

The revenue managers who thrive won’t be the ones who out-spreadsheet the machine. They’ll be the ones who manage more properties, communicate better with owners, and stay open to what the machine can teach them — without losing touch with what their business is actually doing.

Thibault Masson is a leading expert in vacation rental revenue management and dynamic pricing strategies. As Head of Product Marketing at PriceLabs and founder of Rental Scale-Up, Thibault empowers hosts and property managers with actionable insights and data-driven solutions. With over a decade managing luxury rentals in Bali and St. Barths, he is a sought-after industry speaker and prolific content creator, making complex topics simple for global audiences.