AI and Revenue Management: from algorithms to AI agents, what’s really changing

Series: Revenue Management

Dynamic pricing, OTAs, e-distribution, rate disparities, Big Data, RMS, AI, Profit, TrevPAR… The Revenue Management profession is evolving. So are long-held beliefs. So are expectations.

In this series, we take a clear-eyed, pragmatic yet ambitious look at the major trends transforming our industry.

(Re)discover episode one

Wind of Changes · E02 · In brief

Revenue Management did not wait for generative AI to start using algorithms. The real revolution underway is not in data collection (the input), but in the ability to turn analysis into action (the output). AI agents such as Nancie (Revbell) are making Revenue Management more accessible, more affordable and faster to deploy, without replacing reliable data, clear processes or domain expertise. The Garbage In, Garbage Out principle has never been more relevant

From algorithms to AI agents: what has changed and what hasn’t in Revenue Management?

Since the arrival of ChatGPT, Claude and Mistral, artificial intelligence has become an unavoidable topic in the hospitality industry. Some see it as a revolution comparable to the arrival of the Internet. Others are already predicting the gradual disappearance of certain professions, including that of the Revenue Manager.

As is often the case, reality is more nuanced.

Because Revenue Management did not wait for generative artificial intelligence to embrace AI. Long before conversational assistants, Revenue Management Systems (RMS) were already relying on algorithms capable of analysing data, forecasting demand and recommending pricing decisions.

“Is your solution AI-powered?”

This question comes up regularly, whether from hoteliers or Revenue Management professionals. It would be tempting to answer it just as vaguely as it is asked, but that is not Revbell’s vision.

Instead, let’s try to understand what is really changing.

“Revenue Management did not wait for artificial intelligence to start using artificial intelligence.”

Revenue Management was built on algorithms

Since its origins in the 1980s, Revenue Management has been based on a simple idea: using available data to make better decisions. Computing power has increased exponentially, technology has evolved and models have become increasingly sophisticated. Yet the fundamental question remains unchanged:

• For a given date, today, what pricing strategy will maximise the commercial performance of my inventory? Will demand exceed my current price, or will it fall short?

Technologies evolve. The question remains. The tools become more powerful, but their mission stays the same: helping teams better understand demand and make better decisions.

“AI agents represent the next chapter in the history of Revenue Management, not its starting point.”

More data does not always mean better decisions

One of today’s biggest misconceptions is the idea that more data automatically leads to better performance. Reality is, of course, far more complex.

Many data points are correlated. Some are relevant. Others are misleading.

Take the weather as an example. It is generally sunny on the Mediterranean coast in August. Yet August is not peak season because the weather is good. June and July are often just as sunny, but pricing levels are not the same.

Weather is therefore correlated with demand, but it is not necessarily the cause of it. More importantly, a guest booking a holiday six months in advance obviously has no idea what the weather will be like during their stay.

Even if we had a perfectly accurate weather forecast today for 15 August, that alone would not tell us whether the room rate for that night should be increased.

The challenge is therefore not to collect ever more information, but to identify the information that genuinely influences pricing decisions.

That was true yesterday. It will still be true tomorrow.

The real revolution lies in the output, not the input

This is probably the most significant shift brought by this new generation of tools. For years, RMS have already been able to:

• Collect data;
• Validate and qualify data;
• Produce forecasts;
• Generate pricing recommendations.

The real challenge has never been producing analysis alone. More often than not, it lies in understanding that analysis, adopting it and turning it into action.

• How many recommendations have never been implemented?
• How many analyses have never been read?
• How many reports have been produced without generating any real operational impact?

Solutions such as Nancie, Revbell’s Revenue Management AI agent, are helping change this reality.

The goal is no longer simply to generate a recommendation, but to explain it, put it into context and make its implementation easier. This is where the real evolution lies: transforming analysis into action faster than ever before.

“For years, RMS learned how to answer questions. The new generation of assistants is learning how to ask them, explain them and execute them.”

What RMS have enabled for decades and what specialised AI assistants now make possible is not simply access to more data. It is access to more expertise. For independent hoteliers as well as Revenue Directors.

A poor question will always produce a poor answer

The current enthusiasm surrounding these new technologies sometimes overshadows a simple reality.

A tool does not create its own expertise. It relies on the data, business rules and operational context it has access to.

A poorly segmented hotel will remain poorly segmented. The tool may identify inconsistencies, recommend improvements or even suggest radical changes. But it can never fully compensate for weak foundations.

Incorrect data will remain incorrect. Inconsistent assumptions will generate inconsistent recommendations.

The well-known principle of “Garbage In, Garbage Out” has never been more relevant. As tools become increasingly powerful, the consequences of poor-quality input data can spread faster and farther than ever before.

“Garbage In, Garbage Out remains perhaps the most enduring principle in the entire history of Revenue Management.”

Artificial intelligence therefore does not replace the fundamentals of Revenue Management. On the contrary, it makes them even more important.

What real transformations can we expect?

While the fundamentals remain unchanged, several major transformations are already underway. Here are the ones we are already observing in the field, addressing today and designing for tomorrow.

1. Better analysis and better communication

Recommendations are becoming easier to understand, more educational and easier to share across teams.

2. The evolution of the revenue manager’s role

Less time spent collecting data, producing reports and performing repetitive analyses. More time devoted to strategy, decision-making and collaboration with Sales and Marketing teams.

As a result, the Revenue Manager fully reclaims their place within the 4Ps of Marketing: Product, Price, Place and Promotion.

3. The democratization of best practices

Hotels that previously had no access to advanced Revenue Management expertise are gradually benefiting from tools and methodologies that were historically reserved for larger operators. Access to expertise is becoming simpler, faster and more affordable.

4. Lowering the cost of access to Revenue Management

This is probably the most underestimated transformation. Artificial intelligence does not create demand, nor does it increase a hotel’s revenue potential by itself. What it does is enable hotels to unlock that potential more efficiently and at a lower cost: the gains lie as much in efficiency as in performance.

Let’s take the concrete example of outsourced Revenue Management for a 50-room hotel.

With a monthly budget ranging from €700 to €1,300, the market typically offers one weekly Yield Meeting and Revenue Management oversight performed one to three times per week across a 90-day booking horizon.

The most comprehensive outsourced services often range between €1,500 and €2,500 per month.

Today, hoteliers no longer necessarily have to choose between investment and performance. A significant share of this expertise is now accessible through modern RMS, their automation capabilities and the AI assistants that support them.

For a budget historically associated with a standard outsourced service, hotels can now access a level of service that was once reserved for premium offerings more frequent, more responsive and better tailored to their needs.

“The real revolution is not so much increasing revenue potential as reducing the cost required to unlock it.”

The winners will not be those with the most AI

The recent history of Revenue Management teaches us one essential lesson: the highest-performing organisations have never been those with simply the largest volume of data or the most sophisticated technology.

They are the ones that know how to transform information into decisions.

The new generation of tools accelerates that capability. But it will never replace reliable data, sound methodologies or domain expertise.

The hoteliers who will benefit most from this evolution will be those who have built solid foundations: trustworthy data, clear processes and a coherent Revenue Management strategy.

“The winners will not be those with the most advanced tools, but those who know best how to feed them and manage them.”

Because while technology continues to evolve, one thing remains unchanged: the quality of decisions will always depend on the quality of the questions we ask.

The RMS will live on even if you no longer notice it

One of the most significant changes over the coming years may well be the gradual disappearance of Revenue Management interfaces for a large share of users.

Not because RMS will disappear. But because they will become invisible.

A Revenue Manager will most likely continue to access their RMS on a daily basis. A Hotel Manager, Commercial Director or Marketing Director, however, may simply interact directly with their Revenue Management assistant.

Read an alert. Update a rate. Adjust a pricing indexation. Review a recommendation. Share a summary with the team.

Without ever opening the RMS.

This evolution is made possible in particular by technological standards such as the Model Context Protocol (MCP), which allow different systems to communicate securely with one another.

Yet behind this apparent simplicity, the fundamentals remain exactly the same.

AI assistants will continue to rely on business rules, workflows, historical data, forecasting models and structured datasets.

In other words, the RMS will not disappear. It will gradually become an invisible infrastructure serving its users.

“Tomorrow’s RMS will not be less important. It will simply be less visible.”

In practice

To illustrate this shift, let’s compare the experience of an independent hotelier in 2010 with that of an independent hotelier in 2026 using Revbell.

In 2010, the hotelier typically relied on an on-premise PMS, a handful of standard reports and very limited analytical capabilities.

The transition to dynamic pricing quickly encountered several obstacles:

• Limited access to Revenue Management KPIs;
• The need to build monitoring tools manually;
• Updating prices across multiple systems and extranets;
• A significant operational workload.

Even by dedicating one to two hours per day to Revenue Management, the possibilities remained limited. Limited analysis. Limited Yield Management. Limited ability to adjust pricing strategy. Limited impact on revenue. Today, we observe a very different reality:

• A PMS connected to an RMS and a Channel Manager;
• Dedicated Revenue Management KPIs;
• Automated alerts and pricing recommendations;
• Advanced automation capabilities;
• Strategic analyses generated in just a few clicks.

Access to data, analysis and execution has never been easier.

A daily view of business performance. Regular actions. Minimal operational friction.

And a significant impact on performance, requiring only a few minutes of attention each day.

“The history of Revenue Management is not the story of machines replacing people. It is the story of progressively giving people access to ever more powerful expertise.”