The tank. The submarine. The warplane. The machine gun. Poison gas. The Zeppelin! These game-changing offensive weapons were all developed or introduced during World War I. Not since the stirrup, and not until the atomic bomb, was there such a dramatic and novel set of tools for an aggressive military commander to gain ground quickly and decisively. Rapid advances thus became the conventional view, for a moment at least, as to how the war would go. 

In August 1914, Emperor Wilhelm of Germany promised to his departing troops that they would return before the autumn leaves fell. In the British collective imagination, the war would surely be “over by Christmas.” We now know how that worked out: World War I ended up lasting four years and made casualties of an entire generation of Europeans. 

Despite the offensive innovations, we remember WWI as a war of defense, and ultimately a war of attrition. It turned out that one of these novel armaments, the machine gun, was more useful for defenders than attackers. When combined with an agricultural innovation of the same period, barbed wire, and built atop a device as old as dirt, the humble trench, the resulting weapons system was nearly impregnable. 

In the early days of a new technology, we don’t know how it’s going to be used. Over the course of 2023, we have seen the birth of a cottage industry of consultants, pundits and bloggers breathlessly explaining how companies can and will use generative AI to transform their businesses, and indeed their industries. In practice so far, the transformation has been incremental, and mainly localized to departmental efficiencies in customer service, engineering, and sales and marketing. While more profound advances are in store, and we see amazing things brewing in our portfolio, it may be that the generative AI intelligentsia are focused in the wrong place. 

The pivotal moment of commerce comes when a company meets its customer, and while generative AI can make companies more efficient, I increasingly think that in that interaction, customers have more to gain from this new tech than companies do. Just like the machine gun, generative AI might be most effectively used against companies rather than in support of them. The superpowered customer, more so than the superpowered company, may upend our predictions as much as a superpowered defense laid waste to the fond military ambitions of WWI generals.

Shifting the balance of power

The balance of power between companies and customers, whether business customers or consumers, has generally been one-sided in favor of the companies. A company is an expert in its domain while in many cases a customer enters that domain infrequently, often simply for the purpose of a single transaction. In many industries, for better or worse, that is where the margin comes from, in the form of maximizing pricing and minimizing the cost (and potentially quality) of delivery and service, as well as unearned customer retention.

Generative AI, with its ability to synthesize large quantities of complex information into digestible, action-oriented summaries, is already quite capable of being a “customer copilot” and equalizing this balance of power — if not, in fact, shifting the balance in favor of the customer, when a company’s frontline workers are of indifferent quality and have not been similarly enabled. 

I have written about how that balance of power will shift in the banking industry, particularly with the advent, some years from now, of autonomous agents, who can work on behalf of the customer, further reducing friction. Retail banking is an acute example, in that its entire foundation relies on friction and opacity, but there are many other industries that also don’t pass the existential test I mention in that piece. What would happen to your business if every current and potential customer woke up tomorrow and did exactly what was in their best interest, right away?

Perfectly efficient shopping

In retail, the superpowered customer will exacerbate existing trends away from store-based retail built on commodities and toward “n of 1” experiences built on specialty items. But it will get worse when customers can scan an item with their phone in-store and have it delivered to them at home by whomever is offering it at the lowest all-in price. Kludgy versions of this have existed for a couple of decades but, as with so many things, the 100% foolproof and simple version will be 10X better than the 80% correct version. I could easily imagine retailers moving to a “no phones” policy, if it weren’t for the allure of being featured on social media. 

E-commerce will be worse, with browser-based generative AI copilots allowing for a level of automation and breadth of searching that makes it a no-brainer to always pay the lowest price. For once, Amazon might be a loser here, in that its overarching promise of convenience and one-stop shopping has led it into the trap of mild price gouging when it can get away with it, and a high dependence on advertising revenue from manufacturers and platform sellers. 

With perfectly efficient shoppers, advertising will be useless and any level of profit taking on price will be punished swiftly. Costco might be the only commodity retail model that makes money in the coming world.

More buyer advantages in SaaS

B2B SaaS software has come to be viewed as the ideal business model, and that may still be true on a relative basis, but this category, too, will have its challenges. 

SaaS software has been on a tear, growing north of 50% per year for this decade so far. Tellingly, the highest growth (north of 300%!) comes from the smallest customers with the least sophistication. Even prior to the popularization of generative AI, there was a crop of companies helping software buyers better manage their software procurement, leading to fewer vendors and lower average contract values (ACVs). 

In a world where software is easier and faster to build, there will be more entrants in each category with “above threshold” levels of quality. The advent of intelligent software agents that can research, test and auto-negotiate every conceivable tool at no incremental cost will destroy every vendor’s win/loss ratios and effectively cap ACVs even for the winners. This will ultimately be true in every category of procurement, shifting power from vendors to buyers across the board.

A transformed doctor’s visit

Healthcare is the industry with perhaps the greatest imbalance between company and customer, or, as they say, provider and patient. 

The patient has few options to independently assess the efficacy of recommended diagnostics, the results of those tests and the medical wisdom of the treatment options proposed net of side effects, not to mention the historical quality track record and relative cost of that treatment with that particular provider. Within a few years, no patient will ever meet with a doctor without their phone on, recording everything said and ingesting every document and image, with generative AI-driven feedback available on each choice that arises during that interaction.

The doctors will have no option but to basically let their AI deal with the patient’s AI, as subjective diversion from the medical AI’s recommendations will have massive legal consequences. As the son of two doctors, I know well that much will be lost in this future state, but given the medical options available to the majority of humanity, much will also be gained.

No more overpaying for high-value services 

Generative AI will superpower customers in high-value service categories, as well. 

Think of the most confusing bills you get as a consumer — that three-page invoice from your auto repair shop for what was meant to be a simple oil change; the shockingly high bill from a plumber, electrician or contractor; or that incomprehensible four-scroll PDF from the vet that has you looking for the “buy now, pay later” option (why does my dog have so many cavities?!).

Run those through a tuned generative AI model and you’ll quickly see that you’ve overpaid for various parts, consumed more labor hours than what the benchmark would suggest (at a higher rate), potentially had unnecessary work done, and also that you should stop feeding your dog candy corn. 

It’s still possible to win in this new age

There are many theories on why the Allies ultimately found their offensive footing and triumphed in WWI.

The popular one in this country is that the US entry into the war was the decisive factor. The more nuanced take, which I find compelling, is that the Allies got their act together in several ways at once. They put in place a single coordinating authority across the various national armies, and decisively concentrated forces on the Western Front (despite Winston Churchill’s constant lobbying for Eastern Front “Hail Marys”). They also skewed the supply chain/resource equation towards themselves through a German port blockade and access to US markets, and finally figured out how to harness two new weapons systems, the warplane and tank, to counter, flank or overfly the formidable trench. 

My next essay on this topic will lay out a more nuanced perspective on how to win in the age of the superpowered customer, but here again WWI has a useful analogy for us. You can choose to try to win in the way I believe the Allies won the war, by having a unified leadership structure, a clear strategy, sufficient resourcing and high levels of innovation — or you could hope for a savior in the form of some exogenous force, like regulation. 

I say join the tanks!