Should You Let AI Negotiate Your Deals?

Originally published by Forbes Business Council.

Every day there is more news (hype, even) about what AI bots can do. And they can do a lot—with use cases across fields like accounting, marketing, engineering and even law.

Now, there are any number of chatbots promising to "negotiate win-win deals." You yourself might think about testing the waters. But even if you could let AI negotiate your deals, should you?

As an almost lifelong consultant, you know my answer is going to be, "It depends." But I’ll share what I think the answer depends on and why.

Feeling The Need To Standardize?

Before ChatGPT captured the world’s attention, large organizations—in an effort to scale up their negotiations (and reduce costs)—believed you had to standardize contracts. In this way, all the negotiations with suppliers, customers, business partners, influencers, etc., could be handled with limited supervision. Of course, as with any kind of standardization, to get that lower cost, you had to give up some things.

Generative AI offers an opportunity to accelerate the standardization of contract templates. Yet it also takes away the need to do that. With AI’s ability to index for context and to make better associations among terms, AI can enable much more customization of contracts, at scale. No more need for the complex taxonomies and "smart" rules you were putting in place on your own.

Bringing Negotiation Into New Realms

But I think both old-school standardization and AI miss what skilled negotiators can contribute to joint problem-solving, which is what effective negotiation is at its core. Modern chatbots can be taught to express, in casual and natural language, specific pre-scripted trade-offs. They can generate and make use of limited menus of possible variations on price, volume, delivery dates, etc. But they have a harder time making the most of real differences between the parties in their resources, priorities and risk and timing preferences.

We all understand that AI chatbots (at least for now) lack real human understanding, empathy, relationship-building skills and more. Effective negotiators embrace these qualities. They also listen for and ask about the interests of their counterpart that remain unstated behind their bargaining positions. They explore a counterpart’s reluctance about an offer, not by offering a concession, but by trying to understand what makes them uneasy. They test whether their counterpart might accept an option they had rejected previously simply because no one had asked the right questions.

AI models, as good as they are at leveraging patterns detected across a nearly unfathomable collection of reported deals, cannot match an experienced negotiator at solving for a unique hierarchy of complex, nuanced interests.

Therefore, for lower value, lower variability contracts where human negotiators are not likely to have the authority to get creative (and that have minimal payoff), I find that chatbots can do well. Walmart, for example, has found that for negotiating with the many tail-end suppliers for indirect goods (items used in stores, but not sold to customers, like shopping carts), AI can do the job. And since Walmart would not normally negotiate with these suppliers, AI negotiators can extract some value—certainly better than not negotiating at all.

Even some law firms are experimenting with enabling AI to negotiate simple non-disclosure agreements. But I would note that Cisco figured out how to spend almost no time negotiating NDAs at least 15 years ago by putting a standard form up on a website that allowed only a limited set of changes.

Putting AI To Work: Negotiation Preparation

We already know your negotiations are only as good as your preparation, and that’s where AI can really be game-changing. The chatbot doesn’t have to understand, engage with and problem-solve directly with the person on the other side. Instead, it can save time and add enormous value in helping the negotiator prepare.

Gathering and utilizing data to understand market trends, forecasts and risks of various kinds, a well-trained AI model can equip a negotiator to be creative and persuasive. For example, there’s a pilot project underway at MIT designed to provide procurement negotiators with practical and real-time information to create an effective negotiation strategy. The "data scientist in their pocket" makes it possible for them to gauge their supplier’s interests and what their respective alternatives might be—then create different options and offer market data to support them.

I think that training negotiators to use AI is critical. If your questions are all about price, for example, that’s all the analysis you’ll get. But as every corporate procurement department knows, there are more valuable options to explore than whether a supplier will charge less for the same thing.

Preparing well with AI requires not just making sure it has access to the right data but that procurement negotiators are engaging effectively with it through a structured preparation and deal review process.

Where To Start

For more detail on our negotiation experience, visit our Strategic Negotiations page.