Getting started with Sentinel AI: put the RMS to work in your first week

You just activated Sentinel AI inside R2 OS and the question is obvious: now what? This guide walks you through your first week: understanding your starting point, setting your safety net and leaving the system generating price recommendations that you approve.

Why there is no implementation project

A traditional RMS starts with weeks of imports, mappings and files. Sentinel AI does not: it lives inside R2 OS, the hotel operating system where your PMS, channel manager and booking engine already run. The RMS reads the bookings, rates and categories that already exist. Your first week is not about technical setup: it is about learning how the system thinks.

Day one: your starting point and your safety net

Start with what you already have and with what you never want to happen. Price limits (floor and ceiling) are your safety net: no recommendation jumps over them, not even in automatic mode.

1. Confirm your data lives in R2 OS: complete room categories, current rates and recent bookings. They are the raw material of the forecast, and there is nothing to import.
2. Set your floor per season and category: the minimum price you would accept for that room, even in your worst week.
3. Set your ceiling: the maximum you can defend on your strongest dates, the point where your guest would say “too much”.
4. If your seasons differ a lot from each other, repeat floor and ceiling for each one. A few minutes today spare you surprises later.

Days two and three: look outward and look ahead

With the net in place, it is time to understand the context: what your competition charges and where your demand is heading.

1. Build your competitive set in the rate shopper: comparable properties from Booking and Airbnb, same area, similar category and the same type of guest. Four to six is usually enough (illustrative example).
2. Review what you see for each competitor: price, occupancy and reviews. Vacation rentals compete with you even when they look nothing like your hotel.
3. Open your forecast and locate three numbers: OTB (what you already have on the books for a date), pickup (the new bookings that came in during a period) and pace (whether you are ahead or behind your historical curve).
4. Ask for the why behind any figure: Sentinel AI shows you the method, the sample and the events behind each projection. And if it lacks history to project, it tells you so, it does not invent.

The rest of the week: from data to decision

This is where the RMS starts paying for itself. Every day you will receive price recommendations with their explanation, and each one is born pending: nothing changes until you approve it.

1. Read the why before you look at the price: pace ahead, strong pickup, an event on the calendar, a move in your competitive set.
2. Approve the ones that make sense with one click and reject the ones that do not. Rejecting also trains your judgment: note why you disagreed.
3. Compare each recommendation against what you would have done. That difference is the value of the system, or the signal that your limits need adjusting.
4. Once you trust the pattern, decide whether to turn on auto-apply. Recommendations will apply on their own, always inside the floor and ceiling you defined, and everything stays audited.

Common mistakes in the first week

• Turning on auto-apply on day one, before reviewing a single recommendation.
• Leaving the floor too low “just in case”: the floor is your last line of defense, not your panic rate.
• Building the competitive set on aspiration instead of reality: compare yourself with whoever competes for your same guest, not with the hotel you would like to be.
• Ignoring the why behind recommendations and looking only at the number. The explanation is where the learning lives.
• Expecting perfect results in seven days: the forecast improves as the system accumulates more of your own history.

What it looks like in Sentinel AI

Everything above lives in one place inside R2 OS: your forecast with pace and pickup, your competitive set, your limits and your pending recommendations, every number with its why one click away. No imports, no spreadsheets and no black box.