Addressing EV Charging Pain Points with Artificial Intelligence

Subscribe To Download This Insight

By James Hodgson | 2Q 2024 | IN-7368

Engaging with public charging infrastructure is a major pain point for consumers, and could stunt future growth of Electric Vehicle (EV) adoption. Google’s forthcoming additions to the Maps application give important insights into the role that Artificial Intelligence (AI) could play in easing the use of existing infrastructure, with even more value potential to be unlocked through deeper integration with automotive datasets.

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.


Leveraging AI to Drive EV


In April 2024, Google announced a number of upcoming additions to the Google Maps application designed to improve the experience for Electric Vehicle (EV) drivers. Alongside some augmentations to map content and visualization in order to better highlight local chargers and their charging power, Google has also developed a compelling generative Artificial Intelligence (AI) use case for the automotive market. By reading and summarizing the content of user reviews of different public charging stations, Google hopes to address a common consumer pain point—navigating the fragmented and unreliable public charger network to quickly find the best charge point for your location or route.

As the EV trend continues, the public charging network will come under greater scrutiny, as more consumers without access to private charging come to rely more on public charging. Even for occasional Direct Current (DC) charging, the next wave of EV consumers will not have the same patience as the tech enthusiast or first adopters that have defined EV ownership, to date, with the current charging experience. More generally, continued growth in EV adoption will rely on a suite of services that support consumers in getting the best out of their EVs, and Google’s additions to the Maps application offer important insights into the role that AI can play.

The Major EV Charging Pain Points


In almost all regions, the public charging network is in a sorry state. As a consequence of poor planning, a deeply fragmented ecosystem, and poor maintenance, consumers regularly face pain points that are often elaborated in user reviews. It is, therefore, somewhat ironic that these reviews, combined with the power of generative AI, could prove the most potent tool developed, so far, to address these pain points.

  • Difficult to Find: While conventional fueling stations tend to be prominently positioned, charging stations tend to be collocated with shopping centers or within parking structures, and are often deliberately positioned out of the way to prevent Internal Combustion Engine (ICE) owners from parking in these designated parking spots. This can cause a lot of frustration, as EV drivers, often with very low battery, roam around the local area to find the exact location of the charging station. Often, instructions on how to find these chargers are left by users in reviews posted in various apps, including in the form of Google Reviews. Leveraging generative AI can help with reading the content of these reviews, identifying the salient points giving useful directions, and then summarizing this information in a very digestible way.
  • Inaccurate Power Rating: EV charging stations, with the exception of Tesla’s Superchargers, are constructed with components from an array of fragmented suppliers, any one of which can become a source of failure. Even if most of the components such as the transformers can support a certain fast charging rate, if the cable supplier has not provided a cable of the appropriate gauge, the charging rate will be limited, and often less than advertised. Once again, consumers tend to record actual charging rates in reviews that can be automatically and consistently ingested through generative AI, which could, in turn, help give consumers more accurate expectations for charging times.
  • Broken Infrastructure: It is not uncommon for 20% of the EV charging network to be non-functional at any given time in some regions. Due to historic fragmentation, charging standards such as CHAdeMO, CCS, and NACS are in concurrent operation today, and charging stations will feature plugs corresponding to some or all of these standards. Issues at charging stations may, therefore, only impact one of these charging standards, leaving others unaffected. Again, the ability of generative AI to read and understand user reviews, associating consumer complaints with specific standards can help direct EV drivers to the best charging points depending on the specifications of their own vehicle.

At the foundation of most EV charging frustrations is the extraordinary fragmentation of the charging ecosystem, with consumer reviews representing the most useful nexus point of this fragmentation, and generative AI the best tool to consistently turn those reviews into helpful guidance for consumers.


Automotive Data Must Compensate for Infrastructure Failures


While EV reviews can hold a wealth of useful insights to be aggregated, the human in the loop creates a point of failure. As mentioned above, the typical profile of an early EV owner has been that of an enthusiastic first tech adopter, with an active community sharing experiences and recommendations for best use of a novel technology. In many cases, these consumer perspectives have made up for a failure of Charging Point Operators (CPOs) to monitor and maintain their charging infrastructure, or to provide consistent and reliable experiences with respect to charging rates and payment. While some of this can be attributed to the teething problems of a novel technology, the failings have persisted to this day as the lines of responsibility for maintenance remain blurred, and the incentives to perform maintenance remain low.

In the future, shortcomings in insights from the infrastructure will be plugged, not only by consumer review data, but also with data generated by connected EVs, with even basic patterns of low-volume data providing very useful insights into the status of charging infrastructure. Even if a certain charging station cannot report its status, a series of vehicles with a low state of charging entering the vicinity and leaving within a brief time window would be a useful indicator that this equipment is not functioning. Even if a charging station is inaccurately listed as supporting a fast charging rate, a succession of models remaining at the charging site longer than would be expected to achieve an 80% state of charge can be a useful indicator of the actual capabilities of that charging station. Both of these examples could be enabled with a combination of simple probe data and battery management data, making for an easy business case.

While automakers might, with some justification, point the finger at the charging ecosystem for most consumer frustrations with EVs, they must be proactive in working to solve these pain points in order for consumers to have an enjoyable EV experience that they then associate with the automaker’s brand.


Companies Mentioned