Preparing for the exam

A Chengdu software company opens its first overseas research centre in Bangalore, India. Before the launch, headquarters hands the new site director a one-page “India playbook” built from Hofstede and GLOBE scores: Indian staff expect top-down direction, respect seniority, and prefer managers to decide rather than discuss. The director follows it closely. Senior engineers set the tasks, junior staff are rarely asked for input, and promotions track years of service.

Within eight months, 30% of the team has left. Most who leave are in their late twenties; many hold master’s degrees from abroad, and several join local startups that “let us own the work.” One team lead writes that the playbook “describes my parents’ workplace, not mine.” The Bangalore HR manager notes that the local engineering market is young, mobile, and fiercely competitive.

Diagnose what has gone wrong, and advise headquarters on how to use cultural frameworks here.

A Chinese appliance maker acquires a mid-sized German engineering firm and runs it as a semi-independent unit. To keep collaboration fair, the joint steering meetings are formally bilingual: every agenda, slide, and decision is recorded in both Chinese and English, with interpreters present.

In practice, the Chinese managers hold a short call in Chinese the evening before each steering meeting. By the time the formal meeting opens, the main decisions are already aligned among them. The German engineers notice that their questions rarely change the outcome; the meeting confirms a position rather than forming one. Over a year, three senior German staff leave, and a survey shows the acquired unit feels “consulted, not involved.”

The headquarters project lead is puzzled: the language rule is balanced, the interpreters are good, and nothing is hidden.

Diagnose why a fair, bilingual rule still produces exclusion, and recommend how to redesign where decisions are actually made.

A multinational rolls out an AI tool that writes and translates summaries of every cross-site meeting, so teams in different countries can follow decisions made elsewhere. Headquarters likes it: the summaries are fast, and a check confirms they are factually accurate.

Two complaints arrive. A summary of a China meeting is read at the German head office. In the meeting, a manager had said the timeline “might be worth looking at again” — a careful way of saying it would not work. The summary is accurate on the facts, but records only that the team discussed the timeline and raised no objection. The soft refusal is gone, and Berlin plans on a commitment that was never made.

A blunt German review of the Chengdu team’s prototype is then summarised and sent to Chengdu. In Germany the review was routine; everyone left content. Rendered into plain Chinese, the same words read as a personal attack, and Chengdu concludes the decision has gone against them.

Diagnose why accurate summaries fail in opposite directions, and recommend where human review must stay.