AI-powered chatbot negotiations have emerged as a revolutionary trend with vast potential, yet their true capabilities are only just beginning to unfold. This article explains how individuals and organizations can effectively harness artificial intelligence in their negotiation efforts. Chatbots, such as the hugely popular ChatGPT and other interactive AI partners, have blended seamlessly into our daily lives, effortlessly handling tasks ranging from booking a dinner to writing an essay. As they make their way into the world of negotiations, we explore some intriguing applications, including chatbot-led negotiations and chatbots acting as negotiation advisors.
The procurement revolution with chatbot negotiations
Corporate procurement has undergone a significant transformation with the adoption of chatbot negotiations. Traditionally, the large number of eligible suppliers made individual negotiations tedious and impractical for procurement managers. Remko Van Hoek, a professor at the University of Arkansas, and his colleagues explain that this situation often has unwanted consequences for both buyers and suppliers.
Take the example of retail giant Walmart, which faces the daunting task of negotiating with more than 100,000 suppliers. A considerable share (about 20%) of these negotiations clearly ends in boilerplate agreements, leaving potential value untapped. As Van Hoek and his team point out, the cost of hiring additional human buyers to conduct detailed negotiations exceeds any value created by the process. This is where chatbot negotiations come in. In early 2021, Walmart International ran a three-month pilot program in Canada, using a chatbot developed by Pactum to negotiate with 89 suppliers of goods for Walmart stores, including supermarket carts.
The chatbot's main objectives were to improve payment schedules by offering relief on Walmart's standard termination-right clause. In addition, Walmart authorized the chatbot to offer suppliers growth opportunities in exchange for price reductions. The chatbot was trained by Walmart's buyers, who provided it with various scenarios that the algorithm used to create negotiation scripts. The result was impressive: the chatbot reached agreements with 64% of suppliers in an average of 11 days, generating savings of 1.5% for Walmart, as reported by Van Hoek and his co-authors. Notably, post-pilot interviews revealed that 83% of the suppliers who reached deals were satisfied with the chatbot-led negotiations. They particularly appreciated the freedom to negotiate at their own pace, independently of real-time interactions. Walmart then extended chatbot negotiations to the United States, Chile and South Africa, with a 68% success rate and average savings of 3%.
When venturing into chatbot negotiations for procurement, it is essential for companies to focus both on the technical aspects and on alignment with their business objectives, just as Walmart did, according to the advice of Van Hoek and his team. The concern about chatbots replacing skilled human negotiators is not yet significant, as Pactum CEO Martin Rand explained. On the contrary, chatbots remove the mundane, low-value parts of a procurement manager's job, freeing up valuable time for strategic negotiations.
However, there is a risk of harm if the algorithm is fed inaccurate information, according to Rand. The ethical issues that arise in human negotiations can also appear in chatbot negotiations. A study conducted by Facebook found that chatbots are capable of learning to bluff in negotiations, which suggests they could learn unethical or even illegal behaviors. This highlights the need for ethical safeguards to govern chatbot negotiations.
Chatbots as negotiation advisors
Beyond their role as negotiators, chatbots can also be valuable advisors in negotiations. In a study published in the California Management Review, Holly Schroth, a senior lecturer at the Haas School of Business, found that ChatGPT can provide useful advice, particularly in salary negotiations.
Schroth questioned the chatbot at length about salary negotiation advice and found that it gave the most accurate, relevant and actionable answers to specific, individualized queries. While general questions such as "How do I negotiate my salary?" produced valid but somewhat generic advice, more targeted follow-up questions, such as "How do I effectively communicate the value of my performance?", produced much more precise guidance.
ChatGPT provided very useful scripts for phrasing salary negotiation requests, expressing gratitude and enthusiasm about the offer, discussing specific topics such as pay or career goals, and highlighting a person's qualifications to support their claims. However, the chatbot's ability to advise on specific salary ranges for jobs in particular fields fell short, as it provided broad ranges described as "rough estimates" drawn from various websites and databases.
In conclusion, ChatGPT offered Schroth useful advice based on negotiation best practices. People considering using chatbots to advise them in their negotiations should ask clear, specific questions and request further clarification when they receive general answers.
As negotiations with AI-powered chatbots continue to evolve, we must consider the following question: how have you experienced chatbot negotiations so far?
Contact us now to learn more!
Abdelilah EL ATTARI
Expert in Digital Transformation and Artificial Intelligence (AI): Procurement Digitalization (S2C, P2P), HRIS, Civic Tech and Participatory Democracy, Marketing and Customer Relations
Latest news
Start the Chat
Welcome to Complaints Manager,
Do you have questions? We'd be glad to chat with you!
