Writing this paper solidified something in my brain: we live in the world of Fewer Handshakes and More Algorithms.
Remember the times when we used to bargain with vendors in the street markets for 3 bucks? Well the poor vendors we were arguing with had that same exact conversation with at least 30 people that day and they’d argue with 30 more after you’re gone, the man is running on pure spite and the irritability meter is high, now get this, your opponent is a chatbot that never gets tired, never blinks, and has read every deal your side ever made. Welcome to the world of bargaining AI.
Large companies used to negotiate with only about 20% of their suppliers, simply because humans run out of hours, and also because no one is getting paid enough to yell through the screen every single day. Then AI negotiators like Pactum, Luminance, Nibble and Keelvar showed up and started talking to thousands of suppliers at once. They read old contracts, track market prices, learn each supplier’s negotiating style, and run multiple rounds of offers and counteroffers without anyone lifting a finger. Walmart’s bot closed deals with 64% of its targeted tail suppliers in about 11 days, Vodafone trimmed procurement costs by about 11%, and the paper reports cost reductions of up to 40% overall, with cycle times cut by 20 to 50%. Maersk, Sanofi, Wesco and BT have all swapped manual haggling for bots, freeing their human teams to do the strategic work.
But humans are biased creatures and fairness is in the eye of the beneficiary. When a decision goes your way, you rarely care whether a human or an AI made it. When it goes against you, you start scrutinizing the decision-maker, and AI often gets labeled as the more impartial one, the irony isn’t lost on anyone here. That perception wobbles the moment people learn the algorithm was trained on biased human judgments. Workers start shaking in their human boots because the very thing that was supposed to make their life easier is now aiming to take over it completely. AI can threaten professional identity, the sense of “this is what I do and why I matter.” The paper links it to lost autonomy, job insecurity, technostress and the feeling of being reduced to a data point, especially since bargaining was long considered a uniquely human skill.
But don’t make a face just yet it gets worse, the companies gaining the most from this tech (Walmart, Maersk, Sanofi, BT) also share histories of bribery cases, labour disputes or regulatory fines, because when has a big corporation ever treated their workers fairly? Small suppliers, who lack the data and staff to push back, get squeezed into thinner margins and stay in the contract because the alternative is being dropped. The power imbalance is very prominent here.
Where is it headed? Toward AI-to-AI negotiations, where bots on both sides settle deals while the humans sip coffee, and eventually into sales, HR and management or into little bunkers going extinct, or at least that’s what the doomsday youtuber on my screen says. Bargaining AI would likely thrive rather than fail, but it needs transparency, governance and real human oversight, because an AI is only as fair as the data and design behind it. The takeaway: the machines are very good at getting a better deal. The open question is for whom, and whether the people on the other side of the table (and the ones whose jobs this changes) get a say in it.
And if this still hasn’t helped you win a haggle at the vegetable market, remember that the vendor is not yet a chatbot. Use that advantage while it lasts.