VN
ENIn the hospitality industry, one of the most challenging problems is not knowing what customers need today, but rather predicting what customers will need in the coming days, weeks, or months. A hotel can see current booking numbers, occupancy rates, revenue, or past booking behaviour, but this data only reflects part of the picture. What businesses truly need is the ability to transform historical data and current market signals into forecasts that can support future business decisions.
This is where AI Predictive Analyticscomes in. This technology uses historical data, customer behaviour data, booking trends, room rates, seasonality, events, weather, source markets, online channel behaviour, and many other market signals to identify recurring patterns, thereby predicting the probability of future demand. This trend is becoming increasingly significant as Vietnam's tourism and hospitality market enters a growth phase, yet simultaneously experiences strong differentiation across destinations, segments, and source markets. According to Savills, Vietnam is projected to welcome 21.1 million international visitorsin 2025, a 20.4% increase year-on-year; the hospitality sector's RevPAR is expected to rise by nearly 15% compared to 2024. Savills also states that Vietnam aims for approximately 25 million international visitors in 2026. This indicates an expanding market demand, but it does not mean every hotel can grow by relying solely on traditional operations. As demand changes rapidly, the ability to forecast demand and make early decisions can become a crucial competitive advantage.
What is AI Predictive Analytics and how does it differ from traditional data analysis?
Predictive Analytics is a method that uses past and present data to forecast future possibilities. When combined with AI and Machine Learning, the system not only performs fixed statistical calculations but can also detect complex correlations within vast amounts of data, continuously update models with new data, and provide real-time forecasts.
For example, a hotel might notice that its occupancy rate in September this year is lower than the same period last year. Traditional analysis might simply compare the two figures and conclude that demand is decreasing. However, a Predictive Analytics system can simultaneously consider the destination's event calendar, flight schedules, market-specific customer sources, booking pace over the last 30 days, competitor pricing, average booking lead time, cancellation rates, and online search behaviour. From this, the system can detect that actual demand has not decreased, but customers are simply tending to book closer to their arrival date. The difference lies in the fact that AI doesn't just answer "what happened?" but addresses the question "what is likely to happen next?". This is a crucial shift from Data Analytics to Predictive Analytics and then to Prescriptive Analytics, where data not only helps businesses foresee the future but also supports recommendations for appropriate actions.
Why is forecasting customer demand increasingly important for hotels?
Hospitality is an industry with a unique characteristic: products cannot be stored, and the revenue from a room lost tonight is almost impossible to recover tomorrow. If a room is not sold tonight, the business cannot "inventory" that night's room to sell on another day. Therefore, demand forecasting is the foundation of Revenue Management. If the forecast is too low, the hotel might reduce prices too early, missing opportunities to maximise ADR and RevPAR. If the forecast is too high, the hotel might hold prices too long, leading to low occupancy and ultimately having to drastically reduce prices to fill inventory.
The reality of the Vietnamese market shows that demand is recovering but unevenly. Savills Q1/2025 report noted hotel occupancy in Hanoi reached 76%, an 11% year-on-year increase, while Ho Chi Minh City reached 68%. By Q2/2025, Savills reported average occupancy in Hanoi at 72%, with an average room rate of approximately 108 USD/night. These figures highlight an important point: the market is not only growing broadly but also fluctuating by city, season, segment, and customer source. A good demand forecasting model therefore cannot rely solely on the hotel's own historical data but needs to incorporate multiple external signals.
How does AI Predictive Analytics forecast customer demand?
Essentially, an AI Predictive Analytics system functions like a "forecasting brain" built upon the hotel's data foundation. The system receives data from PMS, CRS, Booking Engine, CRM, RMS, POS, websites, OTAs, and other market data sources. AI then cleans, standardises, and analyses the data to identify recurring behavioural patterns. One of the most critical data groups is booking data. The system can analyse booking pace by day, lead time, cancellation rate, no-show rate, length of stay, booking source, room type, and the price chosen by the guest. When this data is continuously monitored, AI can detect subtle changes in booking pace before they become clear trends in end-of-month reports.
The second data group is customer behavior. A customer might not have booked yet but continuously visits the website, views a specific room type, checks prices for weekends, or interacts with an advertising campaign. When combined with the behavioural history of similar customer segments, these signals can become input data for the forecasting model. The third data group is market intelligence. This can include event calendars, holidays, peak seasons, new flight routes, destination trends, competitor pricing, weather conditions, or fluctuations in international customer sources. Deloitte also emphasises that one of the major challenges for AI in the Travel industry is that data is scattered across multiple systems, and a significant portion lies outside the business's systems, such as weather, traffic, attraction opening hours, or visitor numbers at various locations. When these data layers are combined, AI can generate forecasts such as: room demand for the next 14 days is likely to increase, which customer segments have a high probability of booking, which room types will be prioritised, the price sensitivity of each segment, and when to increase or decrease marketing budgets.
Can AI predict what customers will buy next?
A further advancement of Predictive Analytics is forecasting not only room demand but also customer demand. For example, a customer who has stayed multiple times at the hotel on weekends, typically books a Deluxe room, uses the spa, and tends to add dinner. If the system detects this person is starting to search for a room for a similar stay, AI can determine the booking probability and simultaneously forecast additional services likely to be used. The hotel then doesn't need to send a generic message like "Book now today." Instead, the system can trigger a more tailored recommendation based on each customer's predicted behaviour, such as a room upgrade, a spa package, a breakfast package, or late check-out. This is the intersection between Predictive Analytics and Personalization. Euromonitor notes that AI is ushering in a new era of hyper-personalisation in Customer Experience, Marketing, and product innovation, thereby creating additional revenue opportunities for brands.
When AI forecasts demand, Marketing also changes
In traditional marketing models, businesses often launch campaigns based on a fixed marketing calendar. For example, a month before summer, a resort campaign would run, or before the April 30th holiday, a holiday package would be launched. While this approach can be effective, it doesn't fully leverage real-time behavioural data.
AI Predictive Analytics allows marketing to shift from calendar-based marketing to demand-based marketing. If AI predicts that demand from Hanoi-based family groups for a coastal resort will increase in the next three weeks, marketing can increase advertising budgets sooner. Conversely, if forecasts show a customer segment with a low conversion probability, the budget can be reallocated to segments with higher propensity. This is particularly important given the rapid changes in Vietnamese digital behaviour. Decision Lab reported that in Q1/2025, approximately 80% of surveyed Vietnamese consumers stated they had used AI tools, indicating that AI is transitioning from a novel technology to a part of daily digital behaviour. Another Decision Lab study in 2025 showed that 78% of online Vietnamese had used AI in the past three months, and 33% use AI daily. As customers increasingly use AI for searching, comparing, and decision-making, hospitality businesses also need to use AI to understand and predict the behaviour of those very customers.
From "analysing the past" to "predicting the future"
The data maturity journey of a hotel can be envisioned across four levels. At the first level, businesses answer the question “What happened?” through reports on occupancy, ADR, RevPAR, revenue, and bookings. At the second level, businesses explore “Why did it happen?” by analysing customer sources, sales channels, booking behaviour, seasonality, and market factors. At the third level, AI Predictive Analytics answers the question “What is likely to happen next?” by forecasting demand, booking pace, cancellations, customer propensity, and revenue. The highest level is “What should the business do?”. This is Prescriptive Analytics, where the system can combine forecasts with business rules to recommend appropriate pricing, inventory, campaigns, staffing, or promotions. This also aligns with the trends in the travel industry. Deloitte suggests that AI is being increasingly applied in Travel, from Customer Service and operational optimisation to predictive maintenance, shopping, and discovery. Meanwhile, Euromonitor notes that AI and automation are transforming the entire travel journey towards a more seamless and personalised experience.
Not every hotel should start with a complex AI system
One common mistake when implementing AI is starting with technology instead of starting with a business problem. Businesses don't necessarily need to immediately build an AI system capable of forecasting hundreds of variables. A more practical approach is to begin with a problem that has clear value, such as forecasting 30-day occupancy, booking pace, cancellations, or demand by customer source. Once the model proves effective, businesses can expand to dynamic pricing, customer propensity, upselling, staffing forecasts, and marketing optimisation. This approach is particularly suitable for small and medium-sized hotels, where technology and data resources are limited. The ultimate goal is not to "have AI," but to make better business decisions thanks to AI.
The biggest challenge lies not in algorithms, but in data and people
While AI Predictive Analytics holds immense potential, this technology is not a "magic box." If historical data contains significant inaccuracies, the model may produce imprecise forecasts. If the market experiences an unprecedented shock not present in past data, AI may also struggle to predict.
Furthermore, the human element remains crucial. Revenue Managers, Marketing Managers, or General Managers need to understand the market context to evaluate the results provided by AI. AI can alert that demand is increasing, but humans need to decide whether the cause stems from a temporary event, a long-term trend, or an anomalous data signal. Deloitte also highlights a core challenge of AI in Travel: the technology is only as good as the data it is fed, while travel industry data is often fragmented across multiple systems and sources. Therefore, the most effective model is not AI replacing humans, but AI augmenting human decision-making capabilities.
AI Predictive Analytics is not just technology, but a new business capability
The most important point to recognise is that AI Predictive Analytics is not merely a data analysis tool. For the hospitality industry, it can become an intelligence layer atop the entire operational system, helping businesses better understand market demand, customer behaviour, and potential future changes. When implemented correctly, AI can help hotels forecast room demand, optimise pricing, coordinate staff, allocate marketing budgets, personalise experiences, and increase revenue from ancillary services. More importantly, AI helps businesses bridge the gap between data and decisions. As Vietnam's hospitality industry continues to grow, Savills forecasts that Vietnam will attract approximately 25 million international visitors by 2026, while the market is increasingly segmented by destination and category. This makes accurate demand forecasting an increasingly critical factor for hotels aiming to optimise business efficiency rather than simply reacting to market trends. Businesses cannot fully control customer demand. But businesses can enhance their ability to predict that demand. And AI Predictive Analytics is one of the key technologies helping the hospitality industry transition from "guessing" to "forecasting," and from "reacting" to "proactive."
NewSun Hospitality: Partnering with Hospitality businesses on their data and AI transformation journey
AI only creates value when linked to a specific business problem. For hotels and resorts, AI implementation needs to start with existing data, operational processes, Revenue Management systems, Sales & Marketing, and the revenue goals of each property. NewSun Hospitality aims to partner with Hospitality businesses in enhancing management capabilities, optimising operations, and applying technology to business activities. From standardising processes and building data systems to applying AI in Revenue, Sales, Marketing, and Customer Experience, the ultimate goal is not to implement technology to "follow trends," but to create faster, more accurate, and revenue-generating decisions. If your business is looking for ways to apply AI in demand forecasting, Revenue Management, or customer experience personalisation, contact NewSun Hospitality to develop a roadmap tailored to your business model and real-world data.
- Hotline: +84 768 68 2913
- Email: dosm@nshm.com.vn
- Zalo OA: https://zalo.me/2452770272256938463 (NewSun Hospitality)
- Website: https://nshm.com.vn/ (NewSun Hospitality)
References
- Deloitte. (2025). 2025 travel industry outlook. Deloitte Insights.
- Deloitte. (2026). 2026 travel industry outlook. Deloitte Insights.
- Decision Lab. (2025). The Connected Consumer Q1 2025 report. Decision Lab.
- Decision Lab. (2025). The State of Consumer AI in Vietnam 2025. Decision Lab.
- Euromonitor International. (2025). AI: A turning point in global travel. Euromonitor International.
- Euromonitor International. (2025). Experience more with GenAI for value creation. Euromonitor International.
- Euromonitor International. (2025). Top five trends in travel. Euromonitor International.
- Savills Vietnam. (2025). Viet Nam real estate market report Q1/2025. Savills Vietnam.
- Savills Vietnam. (2025). Vietnam hospitality and residential markets commence new growth. Savills Vietnam.
- Statista Research Department. (2025). Artificial intelligence (AI) use in hospitality: Statistics & facts. Statista.
- Statista Research Department. (2025). AI solutions that hotel chains are currently using or planning to use worldwide in 2025. Statista.
