Why Generative AI is Helping Industries Work Smarter, Not Harder – and Keep their Employee Headcount: A Case for the Jevons Paradox
By Michael Quigley, Co-Founder, Impel AI
If you’re unfamiliar with something called the Jevons Paradox, now is a great time to learn about it.
Coined by economist William Stanley Jevons in 1865, the paradox occurs when new technologies increase the efficiency of the way a resource is used, which is generally considered a good thing (think fuel efficiency). Nevertheless, as the cost of using the resource goes down, demand for that resource spikes upward, meaning more of the resource is used – not less.
Here’s a great example: Facing the Arab oil embargo of 1973, the U.S. government enacted the Energy Policy and Conservation Act, which set fuel economy standards for new passenger cars. These regulations required roughly doubling average fuel economy to 28 mpg by 1985. These regulations generated a series of technological advances, including the invention of the fuel injector, which would increase fuel economy so much that by 1993, gas prices had dipped below $1 per gallon – a decrease of 62% when compared to 1973 prices (adjusted for inflation).
Yet, much to the chagrin of economists, these increases in fuel efficiency did not lead to an ever-decreasing rate of fuel consumption. Instead, low fuel prices created an opportunity for car manufacturers – and an increasing appetite for consumers – for the large trucks and SUVs that now dominate the vehicle market. Fifty years later, fuel consumption in the U.S. is at an all-time high, and America’s best-selling car is the Ford F-150 with an EPA-estimated 16/22 mpg rating.
In other words, when it comes to resources like oil, the paradox is not necessarily a good thing. However, when it comes to the major technological advancement of our era, such as generative AI (GenAI), the paradox is nothing to be afraid of. In fact, it might put to rest one of the biggest fears surrounding GenAI: that it will replace human workers.
Case in point: car dealerships are starting to utilize GenAI in compelling ways to work smarter and not harder — without having to materially reduce their headcount.
The Jevons Paradox and Car Shopping
The way many car dealerships are deploying GenAI solutions has proven to be an important case study for the Jevons Paradox. GenAI is allowing dealerships to enhance their customer experience as they depend on AI-enabled “co-pilots” to do the work of a dozen or more reps, but without the expected loss of headcount.
For example, today’s AI co-pilots are greatly improving dealership productivity. With generative conversational AI, prospective buyers can get answers to any question, 24/7, about a particular vehicle through chat, email, and SMS text. From vehicle features and pricing to trade-in and financing options, advances in AI are providing answers to even the most complex inquiries – responding instantly to customers at any time of day or night.
A customer might not even be aware they are interacting with AI, as generative AI-powered assistants can answer 95% of the questions consumers have. Additionally, AI doesn’t sleep! This means that the days of emailing a dealership with a question at 9 PM and then waiting until 9 AM for a salesperson to get back to you with an answer are long gone.
GenAI applications are going even further than AI Assistants and are taking customization to new heights. Prospective car buyers who visit a dealership’s website will browse and interact with different vehicles and packages. Behind the curtain, the AI collects information from the various interactions, which is then used to tailor messaging and the vehicle features or highlights that are presented in current and future visits. The descriptions and images of the same car can vary widely because the AI will display the optimal combination of features that will best appeal to each prospective buyer based on their demonstrated behaviors.
According to Cox Automotive, Americans’ satisfaction with car buying is rising, simply because they’re spending less time doing it. Behind the scenes, AI is speeding up these consumer decisions. Beyond the personalization of offers that generative AI presents and the always-on availability of automated customer service agents, the precision and accuracy of information that generative AI delivers helps consumers cut through the clutter. No muss, no fuss makes for happier auto sales for consumers and dealers alike.
Research has shown that when dealerships adopt and deploy AI at scale, they achieve labor and software savings between 20% and 40%, with a corresponding average lift in showroom appointment set rates of 25% to 40%. Seamless and complementary interaction between dealership personnel and the AI enables greater team productivity and efficiency, while delivering an even more concierge-like and personalized customer experience. It also enables dealer teams to focus on higher-order activities and customer relationship-building. That’s why we also observe an increase in team activities that help nurture customers more effectively, such as greater outbound sales activity, more live phone calls, more personalized showroom appointments, an increase in the speed and dollar value of closed deals.
What does the future hold?
If the auto industry is indicative of how the world economy will adopt GenAI, we may be about to bear witness to one of the most striking examples of the Jevons Paradox of all time. If that’s the case, the advent of GenAI will likely not lead to the three-day work week – it will instead lead to a flourishing of frictionless 24/7 consumer experiences and increased accessibility.
Like the AI co-pilots being used in car dealerships, there are several other examples of verticalized GenAI co-pilots have already begun to abound:
- Legal co-pilots like Harvey are blunting the dullness of redlining documents for attorneys, and increasing access to quality legal services for all
- Healthcare co-pilots like Highmark are evolving the ways doctors diagnose, and engage with their patients leading to better health outcomes
- Investment co-pilots like Hebbia are helping private equity and venture capital investors manage their deal flow and write investment memoranda
Sixty years after Jevons’ discovery, industrial engineer Allan Mogensen summed up his workplace efficiency process as: “Work smarter, not harder.” AI-enabled co-pilots are proof that this can be done.
The Verticalization of GenAI
As with any massive technology upgrade cycle, such as the one we’re experiencing with generative AI, the question of who will benefit the most inevitably arises: will it be the startups or the incumbents who capture this new market and reign supreme?
The answer to this question is: it depends. Initially, it seems that verticalized players that are delivering the following benefits are most strongly positioned to capitalize on generative AI:
- Proprietary data and industry-specific integrations
- Industry-specific LLM development used in conjunction with foundational models
- Acquisitions of services and technology companies with large installed bases
The size of the global software market is $500 billion per year and is still dwarfed by the global services market, which is estimated to be as much as $5 trillion. If generative AI can automate 10% of the global services market— and tech optimists hope this percentage will be much higher— a new market on the order of magnitude of all existing software spend will be created.
No one is better positioned than verticalized generative AI companies to execute this vision. In particular, those that have the proprietary data, integrations, and conviction to execute transformational M&A will be the ones that endure.
The Jevons Paradox might have been a negative thing in the era of coal and oil, as it meant that the more we tried to cut down on consumption the more we consumed. But in the age of AI, the paradox is nothing to fear. The more efficient we become, the more we can all prosper.
