Introduction

For years, artificial intelligence in hospitality has primarily helped hotels understand what happened. It has powered reporting, forecasting, segmentation, and analysis, helping teams identify trends and make more informed decisions. Dashboards got smarter, forecasts got sharper, and segmentation became more granular. But the work of translating that knowledge into action has remained, for the most part, a human one.

That is now changing.

The next evolution of AI is changing the role technology plays in hospitality. AI agents are moving the industry from insight generation to action execution: identifying opportunities, determining the right response, and helping deliver personalized revenue opportunities in real time, without waiting for a human to interpret a report and decide what to do about it.

As Erik Tengen, President, Hospitality Upsell at Plusgrade, explains:

"The shift is from AI that tells you what happened to AI that decides what happens next."

This shift marks a move from AI as an analytical tool to AI as an active participant in revenue optimization and guest engagement. It is a meaningful change in how technology fits into the day-to-day running of a hotel, and it raises new questions about where automation should lead and where human judgement should still come first.

From insights to action

Traditional AI tools have helped hotels answer questions such as:

Which guests are most likely to upgrade?
Which offers perform best?
When does demand increase?
Which segments represent the highest value?

These are useful questions, and answering them well has already improved how hotels plan and price. But the insights generated by these tools still often require human teams to interpret the data, decide what action to take, and execute the next step. A report that flags a likely upgrade opportunity is only useful if someone reads it, understands it, and acts on it before the moment passes.

AI agents introduce a different approach. Rather than simply surfacing information for a person to act on, they can analyze context, recognize opportunities, and support decisions around:

The right offer
The right price
The right timing
The right guest

Crucially, the goal is not simply more automation for its own sake. Plenty of hotels already automate guest communications in ways that feel generic or poorly timed. The goal of AI agents is faster, more relevant decision-making, the kind that a skilled revenue manager might make if they had perfect visibility into every guest, every offer, and every moment of demand, all at once, across an entire portfolio of properties.

The intelligence behind AI-driven revenue decisions

For hospitality, effective AI is not just about technology capability. Plenty of systems can process data quickly. What sets a genuinely useful AI agent apart is the intelligence behind the system, meaning its ability to learn, adapt, and exercise something close to judgement.

As Tengen notes, AI-driven upselling depends on learning from real conversion patterns and understanding when to act, but equally importantly, when not to. This is a subtle but important point. It would be easy to build a system that simply offers an upgrade to every guest at every opportunity. It is much harder, and much more valuable, to build one that recognizes when an offer is unlikely to land and holds back.

This distinction matters more than it might first appear. A poorly timed offer can create friction, frustrating a guest who has just arrived after a long journey or interrupting a moment when they are focused on something else entirely. A relevant offer delivered at the right moment, by contrast, can improve both guest experience and commercial performance at the same time, turning what could be an annoyance into something the guest actually appreciates.

AI agents allow hotels to move away from static rules, the kind of blanket policies that treat every guest the same way, and towards more dynamic, context-aware engagement that adjusts based on who the guest is, what they have already shown interest in, and what is happening in the property at that moment.

The pre-arrival opportunity

One of the strongest opportunities for AI-driven automation is the period between booking and arrival. During this window, guests are actively planning their stay and considering ways to enhance their experience, whether that means a room upgrade, an early check-in, a spa package, or something else entirely.

This is also a period where hotels have historically struggled to engage guests effectively. Pre-arrival emails are often generic, sent on a fixed schedule regardless of who the guest is or what they might want. AI agents can help hotels move past this by identifying:

Which guests are most likely to respond
Which offers match individual preferences
When demand conditions create opportunities
Which pricing approach is most effective

Rather than sending identical upgrade messages to every guest on the same day before arrival, hotels can move towards personalized offers based on guest intent and behaviour, drawing on signals like past stays, booking patterns, and how a guest has engaged with previous communications. The result is a pre-arrival experience that feels considered rather than mass-produced, while also giving hotels a meaningfully better shot at converting interest into revenue.

Unlocking value during the stay

AI-driven engagement does not stop before arrival. Guest preferences and buying decisions can change throughout the stay, often in ways that are difficult to predict in advance. A guest who declines an upgrade before arrival may become interested later, once they have settled in or experienced the property and developed a clearer sense of what would improve their visit.

As Tengen highlights, static offers often miss these changing moments of guest intent. A single pre-arrival email cannot account for the guest who, on day two of their stay, suddenly decides a better view or a larger room is worth paying for. AI agents can help hotels recognize these signals as they emerge and create more timely opportunities throughout the guest journey, rather than treating the offer process as something that happens once and is then finished.

This matters because the in-stay period is often underused from a revenue perspective, even though guests are, in many ways, more engaged with the property and more able to evaluate what additional experiences might be worth to them.

Connecting revenue and guest experience

Historically, revenue management and guest engagement have often operated separately, both as teams and as systems.

Revenue teams focus on:

Pricing
Demand
Inventory

Guest experience teams focus on:

Service
Personalization
Satisfaction

These two functions have always been connected in principle, since a poorly priced offer can damage guest satisfaction just as easily as poor service can, but they have rarely been connected in practice. Different teams, different tools, and different metrics have made it difficult to bring commercial and experiential thinking together in real time.

AI agents have the potential to connect these functions by combining commercial insights with guest context, allowing a single system to weigh both the revenue opportunity and the guest experience implications of any given offer. The result is a shift from optimizing individual transactions towards improving total guest value across the entire stay.

As Tengen points out, the opportunity is not only increasing ADR, but growing TRevPAR by better understanding the full revenue potential of each interaction, from the initial booking through to checkout and beyond.

What happens next

Over the next 12 to 24 months, AI adoption will likely be less about replacing human decision-making and more about identifying where automation creates the greatest impact, and where it does not.

Some decisions are already well suited to AI-driven execution:

Upsell recommendations
Dynamic offer timing
Personalized promotions
Revenue optimization

These are largely pattern-recognition tasks at scale, the kind of work where AI agents can process far more signals, far more quickly, than a human team realistically could.

Others will continue to require human judgement, empathy, and service expertise. Resolving a guest complaint, handling a sensitive situation, or making a judgement call that depends on reading a room rather than reading data are not tasks that are likely to be handed over to automation any time soon, nor should they be.

The future is not AI replacing hospitality teams. It is AI helping teams focus their expertise where it matters most, taking over the high-volume, pattern-based decisions so that people can spend more of their time on the moments that genuinely benefit from a human touch.

The foundation for the AI agent era

AI agents become more effective with every interaction. The more connected the data, guest behavior, and operational signals become, the more effectively AI can identify opportunities and support decisions, building a feedback loop that improves over time rather than staying static.

Hotels that begin building this foundation today, by connecting their data systems and rethinking how revenue and guest experience teams work together, will be better positioned as AI evolves from a support tool into a commercial engine.

The next generation of hospitality AI will not just predict what guests want. It will help hotels act on those opportunities at the right moment, turning insight into action without the delay that has traditionally separated the two.

AI Agents in Hospitality