IA on the job: negotiating without adopting the other side’s frame?
A supplier announces a price increase, and a manager asks AI to prepare a response: the incoming proposal can easily become the centre of the discussion. The management question is specific: how can an assistant help prepare a counteroffer without making the other side’s terms the implicit benchmark for an agreement?
What the literature establishes
Adam D. Galinsky and Thomas Mussweiler (2001, Journal of Personality and Social Psychology) show, in negotiation experiments, that first offers can act as anchors and influence the agreements reached. They also show that this advantage can disappear when recipients focus on information inconsistent with the anchor, including the other party’s alternatives or their own target. The issue is therefore not simply how to respond to an offer, but which reference points to use when evaluating it.
The frame matters as much as the wording
Amos Tversky and Daniel Kahneman (1981, Science) show that different formulations of equivalent decision problems can change expressed preferences. Their experiments concern neither AI nor, directly, preparation for a commercial negotiation. They nevertheless help distinguish two activities: improving the wording of a counteroffer and examining whether the problem has been presented as a loss to limit or a gain to secure.
The premature conclusion
One might conclude that asking AI for a more ambitious counteroffer is sufficient. That would confuse changing the amount with changing the reference point: a concession calculated against the requested increase remains organised around that request. A different approach to preparation also examines alternatives, service terms, volumes and deadlines, without assuming that all these elements are negotiable. (our executive and employee training programmes)
What the evidence does not establish
These studies do not demonstrate that a generative assistant strengthens anchoring or helps users escape it. Nor do they predict the outcome of a lasting business relationship, where contractual constraints and bargaining power also affect the agreement and its implementation. A better-structured counteroffer remains an outcome of preparation, not proof of a better negotiation.
A practical check in Lugano
In an “IA on the job” workshop in Lugano, an Italian-speaking team from a bank, a fashion business, a trading company or a family SME can work on a fictional supplier request written in Italian. Using the same tool and case file, the team prepares one response by simply requesting a counteroffer, then another by explicitly requesting reference points independent of the incoming offer and non-price elements to examine. The proposed measure is the proportion of negotiating proposals justified by a verifiable item in the case file that is independent of the amount requested; a reference invented by the assistant does not count. Someone who does not know which instruction was used classifies the proposals according to this rule: the comparison tests preparation, not future commercial gains. To go further: explore the AI on the Job training in Lugano, or browse our executive and employee training programmes in Switzerland.
In pictures: AI on the Job in Lugano
- ai on the job
- Lugano
- research
- training lugano