The new business revenue model for rental car operations boils down to a simple concept: the right rates for the right rental cars and the right (real) time.

AI tools are not perfect for such tasks, but they already outperform humans at monitoring digital dashboards and will continue to improve.

For rental operators navigating complex travel patterns and signals, the goal is not to replace decision-makers but to equip them with finer data earlier in the travel demand cycle, informing targeted and accurate pricing decisions.

Rental car operators learned about the details of such AI revenue management during the International Car Rental Show, May 13-15 in Grapevine, Texas.

Meeting Demand From The Data

Sanchit Garg, CEO of Rev AI, an India-based revenue management system for car rental companies, showed how AI succeeds best when paired with industry-specific data that provides context and helps predict rental car demand.

“One thing we’ve seen is that demand signals appear before bookings do,” Garg told the audience during his presentation. “You just may not know where to find them. We want to incorporate real car rental data and context as we apply AI to our systems. And if you do that right, it can work for you.”

Traditional revenue management often relies heavily on booking activity as the primary indicator of future demand, but as Garg explained. Demand signals often appear much earlier in the booking cycle.

Before rental operators can determine pricing, they need to put several processes and plans into place.

Garg outlined a three-layer decision process for rental car operations: External demand, fleet strategy, and competitive landscape.

Questions To Identify Rental Car Demand

He first posed several questions to point to an answer:

·     How are bookings doing, and how do they compare to revenue per unit?

·     How has your fleet performed over the last few months and years?

·     How does it compare to the performance of your competitors?

Garg further asked, “If I were to slice and dice this for any of you and for your cities, you would be keen to know, are you building up to solo travelers or for families? What time of the month or year will you see the hottest demand? How should you price for that in anticipation of that demand versus trying to apply generic rules?”

While bookings commonly indicate how demand varies, other signals can better pinpoint how rental operations should be pricing their cars.

“We also see it from a travel search intent standpoint. As we speak, a family on the other side of the Atlantic is looking to book flights to New York this summer and will be searching for them. And we can capture that search intent from among our data partners.”

Search activity may begin more than 40 days before a trip. Flight bookings are often made about 35 days before travel, while hotel reservations may become visible about 3 weeks before arrival.

Data-driven demand forecasts tend to be most accurate within the last 10 days before rental car pick-up updates, with a forecast accuracy of about 90% to 95%. A month out, it would be about 80%-85%. Competitive pricing usually becomes more aggressive in the final one to two weeks before travel.

Factors For A Rental Fleet Strategy

When considering a fleet strategy and how you measure revenue per day, you can see how bookings trend compared to the previous year. That data can be paired against forecasted demand to gauge pricing, supply, and expected revenue per day and revenue per unit.

“So, what is your fleet telling you, and what channels are you using to get your demand?” Garb asked.

Rental fleet operators should consider a specific approach to travel platforms and channels, Garg said, citing OTAs and .com websites as examples.

“How are you using these different channels? Is your current strategy the right one? You may believe that a channel that has worked for the past five years is still the best option. If so, are you well-positioned on that channel? What do your reviews look like?”

Operators should note the performance of each channel. “It’s important that the pricing is put in that context, depending on the location, on how you built the brand, and whether it’s a premium or budget brand.”

Rental operations should set pricing in the context of the channels they want to sell through, he said.

AI has proved more objective in setting rental rates by looking at demand forecasts, thereby avoiding a race to the bottom, Garg added.

How To Monitor Rental Car Competition

In assessing the competitive landscape, rental operations need to evaluate many data points across the wider travel industry, including airlines, hotels, rental car companies, OTAs, transportation, special events such as the FIFA World Cup, weather developments, and historical and geographic demand patterns.

A State of the Car Rental Industry report, based on survey data, indicated 64% of rental car operators tend to follow competitors when setting most of their rental rates.

Rather than automatically matching competitors, an effective revenue management system evaluates demand forecasts, booking pace, and market conditions. If demand remains strong, there may be no reason to lower rates simply because a competitor reduces pricing.

AI does not set prices in isolation, Garg said. “It does not simply compare competitors’ rates and lower prices whenever a competitor does. Instead, it looks at your forecast to determine whether you are on track and whether it is the right time to raise rates. Even if a competitor cuts prices, AI may recommend holding your rate if demand remains strong.”

AI remembers good and bad decisions and can adjust hours or days later as needed, whereas a human team would need to monitor dashboards multiple times a day and piece information together, he added.

“For a revenue management system, which uses machine learning, it is ingesting millions of data points and processing them. It can factor that into a pricing that could help you increase your RPU and push that into your computer reservation system (CRS).”

Machine learning can even detect clusters that indicate what's happening to a particular rental car at a downtown, airport, or off-airport location. “It knows what the ways in which, if you price up or down, it can impact your demand and your conversions at the end of the day.”

Those time savings, in turn, increase the productivity of revenue management employees, who can shift their focus to pricing strategies instead of monitoring and connecting the dots, he said.

While the AI system is not fully autonomous, it can handle predictive pricing inputs, while humans can address unpredictable variables and strategize, Garg said.

“If you ask it a question, it gives you an answer. Ask a different question five minutes later, and it may give you a different answer. But pricing is not probabilistic; it is deterministic. The system must be accurate and consistent, and that is where machine learning comes in. By analyzing large volumes of data, machine learning can help identify the best decision.”

AI revenue management involves a two-way learning dynamic between the rental operation and the system.

Limits can be programmed into the system based on accumulated data and insights, depending on the data's structure, he said. Models can be adapted to local dynamics based on input from the rental car operator.

By tracking each rental vehicle as an individual unit, operators can adjust pricing to maximize profit. Applying that data across the fleet also indicates whether the fleet mix aligns with customer demand and competitive conditions.