When Bots Do The Bargaining

David Jonker and Christopher Koch

A home is the most expensive item most of us will ever buy. That’s why we invest so much time in negotiating the details of the purchase, often with the help of an experienced Realtor. We haggle our hearts out to reach an agreement that satisfies everyone, and we hope for what’s known as a buyer’s market – a market in which there are more sellers than buyers, giving buyers the leverage to drive (or keep) prices low.

By comparison, we perceive bottled water as cheap, even though the markup on water is many times that of any house. In fact, consumers regularly pay $2 for something that costs the manufacturer only five cents to produce, a 40x increase that makes water one of the most heavily marked-up products in the world.

So why don’t we bargain for water (or toilet paper, or other common and inexpensive items) when we know the markup is so high? In large part, it’s because of this simple equation:

[value of negotiating] = [money saved] – ([time to negotiate] x [financial value of our time])

In other words, if time is money, negotiating is only worthwhile when the amount we save by bargaining down the price of something is more than the value we place on our time spent doing it.

It makes sense for someone earning the median US hourly wage of $18/hour to invest, say, 10 hours to research and negotiate five percent off the purchase of a used vehicle with an average price of $20,000. The buyer spent $180 worth of time to save $1,000.

By comparison, imagine that same person trying to negotiate a discount on a $2 bottle of water. We already know the huge markup of 40x cost should leave room for a win-win outcome for both parties. However, a buyer who can persuade the seller to drop the cost by 50% in just five minutes has still spent 1/12 of an hour, or $1.50, to save $1.

So for now, the time it would take to haggle over most purchases makes it not worth the effort.

But artificial intelligence (AI) intermediaries will soon change that.

The value of speed

What takes humans hours or minutes to do, computers can accomplish in seconds or less. Instead of taking five minutes to negotiate a deal individually, or 50 hours to organize a group to negotiate collectively, a human can give a command and let an AI quickly handle the negotiations in the background. And when a negligible “time to negotiate” is plugged into our equation, the result shows why AI-assisted haggling will be common for many products in the future.

To see what that might mean for the future of customer experience, consider a recent demonstration of the Google Assistant AI’s ability to contact a local business to schedule an appointment. The AI placed a phone call to a business that doesn’t have an online booking system, spoke to a human receptionist, responded with pauses and phrasing that mimicked human speech, and parsed the nuances and context of the receptionist’s comments to complete the transaction exactly as a person would. Now imagine eliminating the humans entirely in favor of negotiations between and among AIs that can negotiate at lightning speed.

As more of our mundane shopping moves online, we’ll trust AIs to organize more of it. Eventually, groups of personal AI intermediaries will band together to barter and buy collectively without our intervention, creating economies of scale to ensure the best deals on whatever they’re purchasing for us.

In essence, this will create temporary virtual marketplaces where AIs cooperate with each other to meet a group of human buyers’ needs without involving them in the negotiations. This AI cooperative buying will have the benefit of numbers; 500 AIs negotiating as one will have more influence than each one negotiating individually. Just as critically, the individual AIs will be able to leverage the combined historical and real-time data from each buyer’s purchasing habits and insights into market trends to influence pricing and product availability before they ever begin to negotiate.

Of course, sellers could have their own AIs on the opposite side of the negotiating table, and they wouldn’t necessarily be at a disadvantage. When buyers have delegated to an AI the purchase of a low-volume, highly customized product, or in situations where demand is high but supplies are limited, the buyers’ AI’s may decide to compete with other instead of joining forces, thus temporarily shifting negotiating power to the sellers’ AI’s instead.

Either way, expect AI cooperative buying to complete hundreds or even thousands of simultaneous, automated transactions each day, with participating AIs gathering data all the while. Sellers will need strategies to attract the attention of these ephemeral alliances and present them with winning offers before they make a choice and dissolve again.

SAP worked with more than a dozen industry experts to uncover five trends that will determine the customer experience over the next decade. “The Future Customer Experience: 5 Essential Trends” report examines each of these trends and offers recommendations for how brands should respond now to prepare.


David Jonker

About David Jonker

David Jonker leads the global thought leadership team at SAP, exploring how technology innovation addresses business, economic, and social issues today and tomorrow.

About Christopher Koch

Christopher Koch is the Editorial Director of the SAP Center for Business Insight. He is an experienced publishing professional, researcher, editor, and writer in business, technology, and B2B marketing.