What Businesses Can Learn From Mobility’s Early AI Adoption

AI addresses pivotal challenges for the mobility industry

AI helps mobility providers navigate a wide range of stressors. Transportation combines the heavy, physical world of rail tracks, airplanes, and car engines with the IT necessary to manage fleets and infrastructure in highly complex operations — all while in the service of customers who must be kept safe, and who demand personalized treatment. AI solutions provide a powerful tool to help address and balance the mobility industry’s need to create a top-notch, safe travel experience.

This need for tech-powered solutions is driving above-average investments in AI, particularly in the auto industry. The value of the AI market in the global automotive industry was $8 billion in 2024, according to Towards Automotive 6. The automotive sector is currently investing an average of 3.5% of revenues in AI, compared to 2.2% for all industries, according to an Oliver Wyman analysis.  

Made for mobility, useful elsewhere

Transportation and mobility have long been the source of advances in society and the economy, from standardizing US time zones to developing “agile” and “scrum” working methods that are now the standard for tech development workflows. More recently, airlines developed yield management techniques under which they sell tickets for different prices based on factors such as how close the sale is to the flight date and how many tickets remain. These systems are now widely used in hotels to sell rooms at differing price points to different customers.  

Transportation’s role as a pioneer in technology and standards-setting continues today, and it will be especially strong as AI powers transformative changes in safety, customer interaction, and activities from vehicle design to driving or piloting. Mobility’s unique history of innovation means the sector’s experiences will be applicable to a range of other industries, especially those that deliver safety-critical services.

Prioritize the right use cases

Customer needs serve as the North Star for mobility providers in deploying AI solutions. Many AI applications in aviation systems, for example, are focused on limiting negative impacts on travelers, such as delays and cancelations. Predictive technologies give automatic alerts when an engine needs servicing, and airlines can then route the plane to an airport where the repair can be performed without resulting in delays or cancelations. To further increase maintenance efficiency, airlines have implemented large language models to give mechanics access to the information they need to make repairs as quickly as possible. These models can maintain traceability back to source documentation for maintenance instructions, helping ensure that information is accurate. While these applications may not be overtly consumer-facing, they are ultimately geared toward improving the travel experience and have accordingly been prioritized. The automotive industry has a similar focus on consumers: Infotainment systems that improve the driver experience and assisted driving safety features that support human operators are being developed with AI.

Measure impact

To justify funding for AI and to develop consumer trust, companies need to be able to identify the benefits of the technology for customers and investors. These benefits come in the form of efficiency, sustainability, safety, and customer service metrics.

The mobility industry, like others, still struggles with measuring impact. For autonomous driving, the importance of measuring impact is related to public trust, and companies need to showcase and communicate the safety advantages of AI in driving that may run contrary to widely held perceptions and fears. “Showing the added value of AI is critical for clients to adopt new solutions. Having the right metrics can just help do that,” says Yann Cabaret, CEO of SITA for Aircraft, an air transport communications and IT firm.

Develop standards and collaborate with other companies

Even as airlines compete for customers, they collaborate to develop technologies and effective procedures to prevent repeated incidents and improve safety. This industry-wide approach provides learning opportunities when adopting new technologies. With the right approach, incremental adoption of new technologies does not disrupt safety or pose a threat to public trust. In other parts of the mobility industry, too, thorough testing and experimentation have ensured that consumers are not put at risk.

This sometimes leads to a cautious approach that slows down technology deployment, as when accidents involving autonomous cars lead to a pause in their use. While this can hinder rollout of new technologies, it ensures technologies are deployed safely and with the interests of consumers in mind. Further, through extensive testing and collaboration with other industry players and regulators, these pauses and other barriers to deployment can be reduced, avoiding unnecessary restrictions while still protecting consumers.   

Use AI for customer-centric solutions

Autos

Driving is also becoming more enjoyable because of AI applications, as navigation apps, infotainment, and driver assistance all improve the experience of riding in and operating a car. Investment in AI tech for car ecosystems rose 400% between 2021 and 2023 to reach $1.6 billion 12, according to Oliver Wyman analysis. Drivers can enjoy hands-free operation of some car functions through voice assistants, facial recognition, and gesture control.  

More than 40,000 people lost their lives in auto accidents in the US in 2023 13, and there is significant opportunity to improve safety outcomes. In 2022, nearly 300,000 people were injured or killed in crashes involving distracted driving in the US 14, suggesting that AI technology could play a major role in making roadways safer in the future.

AI already has made some kinds of crashes less likely. Driver assistance systems brake automatically in case of danger and correct lane departures — in each case, reacting to an outside stimulus to protect the driver and passengers. Autonomous cars — in which an in-car AI system makes all driving decisions — are involved in 76% fewer accidents than human drivers in the areas where they are permitted, according to one study.15

And AI-powered traffic monitoring in dense urban areas can help prevent road accidents. In Melbourne, for example, one company deployed a technology that uses AI to analyze images 16 of risky driver behavior inside a vehicle, such as texting, which are then given to the authorities to determine the likelihood of a traffic offense. In the first two years of deployment since 2019, the software firm reported a 22% reduction in fatalities and an 80% reduction in phone use.

Fuel AI systems with the right data

Effective AI implementation requires access to underlying data to train and operate the AI systems. The mobility industry has had several successes in data utilization — such as predictive aircraft maintenance, insurance premiums informed with driver data, and autonomous driving systems and large language models — thanks to its ability to generate vast, unprecedented amounts of data. But some significant barriers still exist, such as effectively gathering, cleaning, and structuring data to be used by AI systems. Other industries facing their own data access challenges can inform their approaches based on mobility’s experience.

Aviation

For decades, aviation has utilized predictive data — such as sensor alerts from the airframe or engines predicting aircraft health issues — to proactively schedule maintenance operations. That data informs engineers and maintenance teams if and where an aircraft has a problem. This can be paired with information about when and where replacement parts are available, allowing maintenance teams to reduce delays, service impacts, human error, and administrative time — all while keeping planes in use for longer.

Historical maintenance and safety data also is used to train large language models that support maintenance technicians. These AI tools help maintenance crews obtain the information and manuals they need, making them more efficient and potentially helping airlines to expand without increasing headcount.

Making data AI-friendly

Despite major successes in utilizing data, mobility still isn’t taking full advantage of the enormous amounts of data it has available. Airlines, for example, often receive information from suppliers and aircraft manufacturers in the form of PDFs. These PDFs, from several third parties in several formats, can be difficult to input into systems and use effectively, requiring time-consuming manual entry to ensure accuracy. Similar processes can be highly time-consuming for IT teams across all transportation providers. “Ninety percent of our work is data cleaning and data structuring,” Jatish Patel, founder and CEO of Flow Labs, a transportation technology company that works to build integrated transportation systems, told SmartBrief in a recent interview. 18 “In transportation, that’s the problem that’s been plaguing the industry for so long.”

All industries face the challenge of converting data into a form that AI systems can digest, so organizations should implement centralized digital processes. AI can help clean and structure inputs so that they flow into a central system and are stored in digital form. The data is then available to train AI systems to carry out and assist in company operations. The challenge of unclean data highlights the need for companies to develop true digital capabilities to capture the power of their data.

Data accessibility challenges for rail providers highlight a barrier for all industries. Train performance data is essential to driving greater efficiencies and unleashing the potential of AI in rail, but low technology integration along the value chain makes it hard for train operators and shippers to share data. There is also a lack of data transparency in the rail sector. Many datasets are held by different players, and they are incorrectly considered commercially sensitive, thus restricting access 19 for other partners. Removing these kinds of barriers will be essential to opening data access and unlocking the full potential of AI adoption.

Other industries

Many other industries will need to break down barriers to take advantage of AI. In the heavy equipment industry, for example, customers usually assemble fleets of machines from a range of equipment manufacturers, but their data is often stuck in silos that have trouble interfacing between different technology systems. Further complications arise when different equipment is used in tandem, such as a tractor manufactured by one company and a planting or chemical application attachment produced by another.

How can these systems better communicate to enhance fleet management? One potential solution is using several AI models as translators between different systems to aggregate and standardize data, allowing previously disconnected systems to communicate and work together through AI. Mobility’s successes — and failures — in data utilization can help the heavy equipment industry and beyond.

Organize and form partnerships for AI

AI operating models require a different approach from traditional technologies, and companies in both mobility and other industries have not yet made the transition to a new way of working. “There is still a lot of confusion about what AI is at the management level,” says Chin Kian Keong, senior consultant at ST Engineering and former chief engineer at Singapore’s Land Transport Authority.

The main challenge is how to harness AI expertise. Fully 96% of employees say that generative AI can benefit their jobs, according to an Oliver Wyman Forum study. 20 But finding exactly how to capture this benefit is not easy, and 57% of employees report receiving insufficient generative AI training from their employer. By 2027, 60% of employees in the wider economy will need reskilling or upskilling on AI 21, so companies need to prepare effective training.

While these challenges are significant, the potential is enormous: Generative AI can save 300 billion hours of work globally each year, the equivalent of roughly two hours per

internal solutions become outdated. The growing complexity of consumer demands has forced many mobility providers to include players from other sectors with technological know-how. Declining rates of car ownership, for example, have encouraged automakers to combine the capabilities of a tech company, an energy company, and a service company to provide alternatives.

Find the right partners

Partnering with an AI company is an efficient way to obtain leading-edge AI technology. For example, if an aviation company is developing an AI-based system to recognize aircraft hull damage using cameras, a partnership might ensure that it always has the latest AI visual identification tech. “With the pace at which tools are being developed, companies risk being left in the dust by not having an external source,” says David Shmoys, director of the Center for Data Science for Enterprise and Society at Cornell University.

The auto industry has in recent years relied on partnerships to blend its expertise with that of tech companies to deliver new generations of voice and navigation technologies. The best and most widespread examples of in-car software are smartphone connection systems, which work with a variety of infotainment systems in many car models.

For companies in other industries, the decision of whether to develop internal solutions or access outside expertise will depend on available resources, system requirements, data privacy, costs, and potential boosts to safety and consumer benefits.

Educational institutions also provide access to the latest research into the integration of technology and mobility systems. One example is battery experiments carried out at Stanford University 22 through a collaboration with the Massachusetts Institute of Technology and the Toyota Research Institute. Scientists trained a machine learning model on several hundred million points of battery charging and discharging data. Their algorithm predicted the number of cycles that each battery would last and thus whether it was suitable for electric vehicles — a process that normally takes a lengthy analysis and slows down development.

Increase organizational efficiency

The rise of AI will also transform operations and the types of skills needed by the workforce. In the case of airport operations, required skills will become more specialized as AI use expands. Passenger-facing employees will need to upgrade customer service skills and operations teams will need new technology skills to interface with AI systems. The industry also will need to think about how to attract talent into this AI-transformed workplace. Collaborating with academic institutions will be an important way to facilitate interest in aviation and recruitment to meet future industry demands and ensure that workers have future-proofed skills.

Other industries

Other industries can learn from the mobility sector’s experience with partnerships to drive AI deployment. Similar to mobility, the integration of advanced AI systems into construction and agriculture equipment could increase efficiency, from fleet management and harvest project planning to weather forecasting. However, most companies in these industries are not set up to develop complex AI systems internally.

Major construction and agriculture equipment players have begun partnering with tech companies for cloud connected services, but direct AI partnerships (particularly with newer companies) are still nascent. As in the mobility industry, these partnerships have significant potential in automation, AI-powered analytics, and advanced digital twins. Companies in this space should start assessing how to forge partnerships with AI leaders, as has been done in the mobility space, to capture the latest technology in their products.

Expanding the role of partnerships will require companies to undergo a fundamental shift from controlling everything in their system to allowing “black boxes” in which they may not fully understand everything that happens in an algorithm. They will have to learn to trust partners and their systems, and to relinquish control over some of the technology.

Other industries also will benefit from the use of AI recommendations to increase workplace efficiency. In industrial sectors, AI can streamline the process of assigning service technicians to repairs by analyzing their history and the nature of the required repair. Equipment recommendation engines can optimize resource utilization and increase customer satisfaction by recommending the most suitable equipment for a given job. Additionally, text digitization tools can distill extensive repair manual collections and schematics into clear walkthroughs for technicians and format email quotes and service reports into succinct and intelligible summaries, improving communication channels, expediting decision-making processes, and making technicians more efficient and effective.

Work with regulators

Aviation in particular shows how regulation can ensure safety while fostering innovation. In the US, aviation incidents are investigated by a range of agencies, including the Federal Aviation Administration. After serious near-misses and accidents, the FAA conducts scene investigations, data analysis, interviews with people involved, and regulatory compliance checks. Based on its findings, the agency then makes safety recommendations or takes legal action.

While regulation is often viewed as cumbersome, a clear regulatory environment supports innovation by giving companies a framework to test, certify, and deploy new technologies. Even for new AI-supported technologies that may not yet have published rules, such as air taxis, the stability of the regulatory environment gives companies the confidence to innovate and build trust, and provides a guide for working with regulators to bring new applications to market.

One area in need of improved regulation is autonomous driving, where 45% of consumers report anxieties about the related technology, according to a global Oliver Wyman Forum survey. Aviation’s clear framework for implementing new tech solutions may provide a model.

While some autonomous vehicles may be safer than human drivers, they haven’t yet met the level of safety achieved in the aviation or rail industries. After several high-profile safety incidents, some self-driving cars have been pulled off the road, potentially slowing innovation. “There’s been progress, but there remain some real procedural, data, and regulatory gaps,” explained Tilly Chang, executive director of the San Francisco County Transportation Authority, on how to ensure autonomous vehicles 23 serve the public interest.

Mobility providers must do more to assuage consumer safety concerns, and stricter regulation can help. Less than 20% of consumers said they would definitely use automated cars, shuttles, or land and air taxis, according to a 17-nation Oliver Wyman Forum survey conducted in June 2024, while 64% said that stricter safety protocols would make them more likely to use autonomous vehicles. Without clear rules for experimentation, companies face risks from costly lawsuits due to data privacy concerns and technology glitches. Yet the regulatory landscape has not been fully developed.

In San Francisco, for instance, autonomous vehicle companies are regulated by the state, not the city. There is also no way to give an autonomous vehicle a traffic ticket without a human driver. The California legislature is considering bills to address these accountability and transparency gaps, highlighting the importance of an improved regulatory framework to allow autonomous vehicles to progress safely.

The absence of a harmonized regulatory framework for collaborative automation in the US leads “most companies to prioritize local safety — which they should — even if it is at the cost of optimizing overall mobility, traffic fluidity, and energy efficiency,” said Alexandre Bayen, professor of engineering and computer science at the University of California, Berkeley.

Regulation — for autonomous driving and other areas — also can ensure that innovation serves the public good. AI technologies employed by mobility companies can help cities and countries achieve their sustainability goals and make mobility services more affordable. As governments work toward hitting their climate targets, AI systems can help make transit systems more efficient through route optimization and traffic management, thereby reducing fuel consumption and emissions from cars. Shared mobility, like ride-hailing or bike-sharing, can benefit from improved efficiencies from AI models, thus resulting in fewer cars on the road.

Businesses and governments should take note of the profitability potential of AI-powered, safe mobility, as 43% of consumers said they would pay a premium for autonomous transit, according to an Oliver Wyman Forum survey.

Other industries

Just as autonomous driving could benefit from a more established regulatory environment, the construction and engineering sectors would benefit from clear rules around safety and standards for AI systems. Currently, the maturity of regulation varies widely among different authorities. European regulators provide more prescriptive rules, but US agencies have been less involved. Uncertain regulatory environments have led industry groups to step in, but clear frameworks are still not in place. The current lack of clarity creates barriers for businesses seeking to unlock the full potential of AI, and improved regulation could help alleviate these challenges.

Mobility has a long history of AI adoption. The industry has embraced new technology while facing stringent demands for performance in safety, cost, and convenience, which have forced the industry to use any means possible to innovate. These factors turned mobility into a technology leader — even in areas where it is not native, such as AI. Other industries — some with smaller scales or fewer direct participants — share many challenges with mobility. If a company faces a challenge in how to approach AI, there is likely an example to be found somewhere in the mobility industry.