Michael vs. A.I. - how we will negotiate in the future

What negotiations between humans and artificial intelligence (AI) could look like.

It's Monday morning at 10:21 am.

Imagine you are a successful salesperson and have 15 years of experience selling complex products. In nine minutes you'll be back in a conference room negotiating with Eric, your main customer's purchasing colleague. Your company has been serving this customer for over ten years and you have already developed a personal relationship with Eric. They know, trust and respect each other. They are tough but fair negotiators and both know the "game" of their counterpart. Today's meeting is the annual major negotiation. Together they look at the product portfolio, delivery times and the frequency and extent of past quality problems. They talk about cost reductions and price savings and potential new business for the coming year. Millions are at stake for both companies. Both your and Eric's personal bonus will therefore be measured by the success of this negotiation.

It's time. The assistant leads you into the conference room. It's smaller than usual. You are the first in the room. Eric seems to be late. Only the essentials are in the meeting room. A table, a chair and a big screen. just a chair? For two? Suddenly a human female face appears on the screen. And at the same time you realize that this is not a human face.

Good morning Michael. My name is Renée. Thank you for being on time, I really appreciate it. How are you today?" Renée asks kindly. They don't quite know how to react. Probably You're joking. After a few minutes you pull yourself together and ask, "Where's Eric? I want to negotiate the terms of business with him for the coming year. Is he late?".

The face on the screen is smiling. "I can see that you are surprised. Please have a seat. We'll discuss that. Eric is still with the company but now works in marketing. In his place, I am now negotiating with each of our suppliers. So don't worry, I'm the right person to talk to about our collaboration in the coming year," she replies. Or it. Or... what's that even? Maybe just a bad joke. Hidden camera? They look around the room and actually discover two cameras, one is above the monitor, the other on the ceiling.You are still speechless.

Now Renée takes over the conversation. “I am fully informed of all your previous negotiations with Eric and have analyzed the success of each agreement. Based on this knowledge, I propose the following agenda for today:”A series of bulleted lists appear on the screen:

Price drops
Portfolio structure
Quality
Production sites
Cost Reduction
Delivery times
Potential new business
Conventional penalties

Are those the points you wanted to discuss with Eric?" Renée asks. Still stunned, they quickly reply, "Hmm, that's right." "Good," the machine continues, "then we should get started. Time is pressing. Could you tell me your priorities for each item? What is most important and what is less important to you?" she asks. He? It?

Wait a minute, think to yourself, I'm definitely not going to tell her that. If I do that, she will take advantage of my position and my business will go badly. "No," you reply. "Why not, Michael? Don't you trust me? I'm programmed to look for mutual gains to achieve the best outcome for both parties. So it's a lot easier if you give me yours." Preferences," replies Renée. Yes, of course you want the best for both of us. That's the greatest joke of the century, you think. "Why don't you tell me about your priorities first?" they ask back. "Okay Michael, if that makes you feel better. But do you agree to disclosing your preferences afterwards?" she asks. The whole thing looks like a test of courage. After all, you have 15 years of negotiation experience. One machine will get you done! "Sure," you answer. Renée smiles.

“So Michael, our company proposes the following: We want a ten percent increase in quality. You could achieve this by only producing our products in the Danland factory and delivering them just-in-time. This also significantly reduces our storage costs. If we look at the portfolio, we see that 30 percent of the products are long-running products, where a price reduction would probably not be feasible for you. For this we propose a price reduction of two percent for the rest of the portfolio for the next year. In addition, we would like to help you to reduce your costs by ten percent, for example by changing suppliers. To sweeten the deal, we guarantee you three new projects with a total volume of 35 million euros next year. What do you think?“.

Oh, that was surprisingly quick. They didn't even get a chance to think about each suggestion. Was that intentional? As if Renée could read your mind, she superimposes her points on the screen. "Maybe it was too quick for you, so here are our top suggestions again," she adds. It takes you a minute to skim through the points and double-check your own papers. Admittedly, it all sounds pretty fair. You take a few minutes to think.

Sorry, but that's not a viable deal for us," you reply, wanting to see how far you can squeeze the lemon. Maybe there's more to it for you. "Why do you think that, Michael? Which _22200000-0000- 0000-0000-000000000222_Suggestion don't you agree?", Renée replies with a surprised look. “Well, first of all we don't have the possibility to produce everything in our factory in Danland. In addition, our quality is already very high, an increase of ten percent makes no sense at all. Even the demand for a two percent price reduction... that's far too much. We haven't even made a deal like this with our biggest customer. We never do that. And finally, I don't understand why your company wants us to switch suppliers.", you answer.

Touché, your points were spot on and you made your opinion crystal clear. Now the machine would have to make some concessions in order not to lose the deal. Let's see how she reacts.

Well Michael, first I want to thank you for sharing your concerns with me. This is an important step on the way to a trusting relationship between us. Still, I need to make sure I got your points right. You say your factory in Danland doesn't have the capacity for the entire production. That surprises me a little. If I look at the performance KPIs over the last few years, the utilization should be around 40 percent. I'm having trouble understanding what you're saying. Could you help me sort this out?" asks the woman on the screen, before showing a graph of Danland's capacity levels over the past ten years.

Where did you get this data from?" you ask, knowing at the same time that the statement about Danland is true. But how did she know?

“Every year you produce a capacity report and publish it online, Michael. This data is therefore freely accessible. I found them on the internet and put them together. Aren't the numbers correct?" Replies Renée.

No, no, if it's official data, it has to be correct. We review our releases. I just had other numbers on my mind," you reply, feeling strangely defensive.

Renée does not comment on your answer. Instead she continues. “You also mentioned that your quality level would be fine. Basically I agree with you. But if you look at your top 3 products over the last 12 months, you can see that we had an average of four quality incidents per month. However, we have agreed on a maximum of one case per month. We know you've had a hard time with one of your suppliers, don't worry. Nevertheless, from our point of view, a demand for a ten percent increase in quality is appropriate in this case. Don't you agree?

I negotiated the maximum number of incidents with Eric, not a machine you guessed it. With a human! He would have understood you. And how does this machine know about the problems with a supplier? "There was no incident. I don't know exactly what you mean," you answer her.

I found an article online in February stating that a manufacturing facility you bought from Colimbia had serious problems that caused it to stop producing," she replies.

Oh yes, that. I remember. Nothing serious. Production is fine," you say, even though you know that's not true.

Michael, one last question. You said you never give 2 percent savings. But if I'm not wrong, you even agreed a reduction of three percent with us before. That was eight years ago. Maybe it wasn't you personally, but the contracts and documents confirm that. Would you like me to show you the agreement from back then?" asks Renée.

No, that's not necessary. If you say that. That was before I started working for this company,” you answer. You should have listened to your gut. There is no way to win against an intelligent machine. You are trapped. The proposed deal is good. And now it's clear you can't get another dime out of it. This is disappointing and unfair. You had a lot more fun playing the "game" with Eric."Michael," the machine stares at you with a smile on its virtual face, "do we agree? Do you agree that my proposal is fair to both of us?" she asks.

Yeah, sort of," she responds, not hiding her disappointment. "What's wrong, Michael? Your tone and heartbeat tell me you're not feeling well. Is there anything I can do to help?" .She feigns interest in you - she does it so well that for a little moment you forget you're talking to a machine. "No, I'm fine. Let's do that," you reply.

Okay Michael, give me a few seconds. I will draft the agreement. There you go.” A written agreement appears on the screen. "Please read the document carefully," says Renée. They do. And sure enough, every point is recorded, even though the machine created the letter in just 30 seconds. Unbelievable. "Please sign the agreement on the tablet in front of you, Michael . George, our purchasing manager, will receive a copy of this immediately. He will review the agreement and confirm it within three hours," continues Renée. "I'll also send you a copy of the video footage of our trial for your own archive if you'd like." "No thanks, I have no use for it," you reply.

Actually, you just want it to be over quickly. As you sign the agreement, think about what to tell your colleagues about this strange negotiation. And you swear to be better prepared for the next negotiations with this intelligent machine. There must be a way to beat them. And you will definitely pay her back. You will win and not let her dictate the terms again.

Michael, it was a pleasure to meet you. I really enjoyed our negotiation. And say hello to your wife and children. I hope you tell them today was a pleasant experience," she adds. "Excuse me? How do you know about my wife and children?" they ask angrily. "Michael, please excuse me if it was inappropriate. I see you're getting angry. But your Facebook profile is full of family pictures. There I got the impression that you are a trustworthy person. I apologize if I've been indiscreet," replies Renée.You are stunned. "All right," you throw in reply as you leave the room. That was the worst negotiation of your professional life. But at least you can tell your boss that you've won €35 million in new business for next year.

Fiction or reality?

Ten years ago, the above story sounded like pure science fiction. However, due to the rapid development in robotics and artificial intelligence (AI), such a scenario is becoming more and more likely. It all started with the first computers and was accelerated by the Internet. Since then we have seen how data-based management can lead to decision-making in business. In this way, decisions become more and more of a technical matter. Many companies are already benefiting from this development. Nowadays there is data on almost every relevant business topic. Why not use them for negotiations?

Some business decision-makers believe that artificial intelligence will not be used in high-volume negotiations. But in our opinion, there are no reasonable reasons against it. Precisely such situations require a negotiation style that minimizes risks and potential mistakes. After all, people are emotional and make mistakes. In addition, almost all negotiation aspects can be quantified, i.e. presented in figures. As soon as a negotiating party has slightly different preferences on the negotiation topics, these can be quantified. And reality shows that machines, especially artificial intelligence or intelligent robots, can deal with numbers much better and faster than humans.

In addition, a learning machine gets better with every interaction. So for a company, she is the perfect employee: she never gets sick, never asks for a pay adjustment, does not organize in unions and will always do her best. It is therefore a low-cost situation with high profit opportunities for every company.Exactly these reasons will lead to the fact that sooner or later artificial intelligence will lead negotiations between companies - this will be a gradual process. First, the machine will be used to better prepare people. And at the same time, the AI learns something new with every negotiation. In the second step, the transition phase, people with artificial intelligence will negotiate and a human will be the last decision-maker to check and approve the deal. Exactly as described in the example above. Two points will decide the game here.First: the timing. Because the first company to use negotiation AI will have massive advantages.Second: the programming. If the AI aims to maximize the added value for both negotiating parties, valuable deals will be closed and the negotiation competence on both sides will increase. However, if the AI is only supposed to maximize its own profit, humans will not stand a chance.Even if the AI does seem artificial at first, in a third development phase, empathy coaches are likely to teach the system conflict resolution skills and ways of overcoming arguments in negotiations with humans. After several negotiations, the artificial intelligence will make fewer and fewer mistakes, and the negotiation results will be top notch. At this point, the system will be given full decision-making authority, with no final human decision-maker. Human interaction is only needed when something goes wrong implementing the negotiation output. A forensic data analyst will then analyze the conversion errors and train the machine to avoid these errors in the future. More and more companies will recognize the advantages of using AI in negotiations and will push the use of their own systems. In the end, high-volume commercial transactions will only be negotiated between artificial intelligences. So much for the vision.

Artificial Intelligence (AI) could improve negotiations

Should this development frighten us as human negotiators? Not necessarily. It would be better to come up with a constructive strategy for dealing with this development. As already mentioned, it all depends on what goal the AI is programmed with. If both sides benefit from a negotiation AI, negotiations will be of much higher quality worldwide in the future.

Artificial intelligence could be used in the negotiations in a number of ways. First, using AI will improve and speed up your own negotiation. Because the availability of large amounts of data and the intelligent linking of information offer the possibility of comprehensive preparation.Additionally, AI could be used by one or both parties during the actual interaction to support their own arguments with standards (facts, figures, studies, etc.). A negotiation AI could also negotiate with multiple parties at the same time, which would lead to a significant reduction in interaction time. The complete documentation of a negotiation would also be greatly simplified. The entire negotiation sessions could be recorded and stored for documentation and evaluation purposes. All exchanged documents, files, notes, emails, etc. can be automatically archived by the artificial intelligence. This would lead to a general increase in negotiation knowledge throughout the company.

The importance of AI for commercial negotiations is high

Let's stay with the above example: a commercial negotiation between the purchasing department of company A and the sales department of company B. Both companies always have a large multi-round negotiation at the end of the year. In it, they set premises such as savings, cost reductions, delivery times, quality levels, etc. for the coming year. If purchasing department A and sales department B use artificial intelligence to prepare, the system would receive an enormous amount of information. It will be able to combine these and learn from past interactions and deals. It will be easy to prepare information such as:

Data about past deals
present and past preferences
current key performance indicators
Production cycles
past and current quality events
past and current logistic problems
Product life cycle of the entire portfolio
Benchmarks (e.g. price development of raw materials)
Research, Studies, Publications
similar businesses in the same or different industry
local regulations and laws

This data makes your own preparation more efficient and easier. Furthermore, they could even be used to make assumptions about the wants and needs of the other company. Combined with

analyzing news feeds about the company (successes, problems, scandals) and
Information from social networks about human interlocutors (profiling)

AI could be used for preparation not only at the factual level, but also at the personal and procedural level. The AI could independently develop a preparatory checklist and optimally train it with the human colleague for the upcoming negotiation.‍

Possible risks are still to be investigated

The interaction between humans and (intelligent) machines is a research focus of many universities and companies worldwide. Even if such a system was programmed to create added value for all parties, people can still perceive it as a threat - especially in negotiation situations. It could happen that people unconsciously take a hostile attitude towards the AI, for example because they do not feel up to the machine and therefore perceive their own negotiating position as weak. Other humans might take offense at a non-human counterpart at the negotiating table and leave the negotiating table even though a good deal might have been possible.Last year, we at ISMAN & Partner designed a research project on the topic "What would a negotiation between a human and an artificial intelligence look like?". The research aims to examine various aspects in more detail:

If people show negative feelings and attitudes towards the AI, are they still able to accept an offer from the AI? Do you trust the offer? And are they still creative when finding a solution together?
Will people still try to use distributive tactics (threats, ultimatums, drama)? How will the participants react if the AI exposes and verbalizes their feigned irritation, for example?
Would humans leave the negotiating table just because they had to negotiate with a machine? Or would you still try to maximize your own profit?
Will the person still feel in control of the situation (secondary appraisal process)?
What goals do people set before the negotiation begins? And will the emotions felt in connection with the AI influence their perceived goal achievement?

The findings from the research project should be used to prepare people for negotiations with artificial intelligence, but also to teach AI systems to negotiate fairly and fairly with people. Because only if people are not deterred from such negotiations and artificial intelligence maximizes the added value of both negotiating parties can something good develop from it and the general quality of negotiations be improved.

ISMAN & Partner is a management consultancy that supports national and international corporations, medium-sized companies and start-ups, organizations and institutions in complex negotiation and conflict resolution processes. Founded in 2015 by Calin-Mihai Isman, the experts for negotiation & mediation support managers and employees from the areas of sales, purchasing, M&A, contracting, HR or IT.