Low Context, High Stakes: Leading Engineering Across Four Cultures in the Age of AI

I lead engineering from Austin, and most weeks the work runs through four countries: the United States, Poland, Ireland and India. The same sentence lands differently in each. In Austin, “this is great” is a polite acknowledgement. In Kraków, nie jest źle, “not bad”, can be genuine praise. In Ireland, “that’s grand” can mean anything from “excellent” to “we have a problem”. In India, a “yes” in a meeting often means “I heard you”, not “I agree”. Most of the friction in distributed engineering is not technical. It is context.

This essay is about that context: what the research says about how these four cultures differ, what the numbers do and do not mean, and why AI makes this old leadership problem both easier and more dangerous.

Two ways to say the same thing

In 1976 the anthropologist Edward T. Hall split cultures into high-context and low-context. In a low-context culture, the message is the words: say it explicitly, repeat it, put it in writing, and the responsibility for being understood sits with the sender. In a high-context culture, much of the meaning lives around the words, in the relationship, the history and what is left unsaid, and the listener is expected to read between the lines.

Erin Meyer’s The Culture Map turned that idea into a scale managers can use. It puts the United States at the extreme low-context end, and India firmly on the high-context side: communication there is layered and relational, and a lot is said by what is not said. Meyer’s published summaries do not place Poland or Ireland by name, so for those two I rely on my own experience. Polish engineering culture is far lower-context than India’s but leans more on relationships and reading the room than the American default. Ireland shares the English-speaking, low-context default on paper, then softens it with understatement, humour and a reluctance to say the hard thing too plainly.

The twist is feedback. Meyer separates how explicitly people communicate from how directly they criticise. She puts most European cultures on the direct side of negative feedback and Americans and the British in the middle. That matches what I see, with one wrinkle: Poland is the outlier in my four. A Polish engineer will tell you in a code review that your design will not scale, without the cushioning. An American manager wraps the same message in two compliments. An Irish colleague might say it is “an interesting approach”. An Indian colleague might not say it in the meeting at all, and raise it later, one to one. The same comment can read as normal in Kraków and as harsh everywhere else, and a concern that was voiced in Ireland or India can go completely unheard in Austin.

What the data says

The most widely used dataset for comparing national cultures comes from Geert Hofstede, whose original research in the late 1960s and 1970s was built on surveys of IBM employees in dozens of countries. It is maintained today by The Culture Factor, the firm behind the Culture Compass. Hofstede’s 6-D model scores each country from 0 to 100 on six dimensions. Here is how the four compare:

Dimension	US	Poland	Ireland	India
Power distance	40	68	28	77
Individualism	60	47	58	24
Achievement and success	62	64	68	56
Uncertainty avoidance	46	93	35	40
Long-term orientation	50	49	51	51
Indulgence	68	29	65	26

The four cultures split along two lines, and agree on two things.

Power distance: two flat cultures and two hierarchical ones

Ireland (28) and the US (40) expect managers to be accessible and to be challenged. Poland (68) and India (77) accept hierarchy as natural: the boss is expected to decide, and contradicting a senior leader in front of the room is costly. India’s score comes with a clear expectation of a directive, protective leader and, in Hofstede’s words, negative feedback that is rarely passed upward. Put an Irish or American executive in front of a Polish or Indian team, ask “any concerns?”, hear nothing, and you have the most dangerous meeting in distributed engineering: everyone leaves believing the plan is sound, and the people closest to the problem never said so.

Uncertainty avoidance: Poland alone

This is the largest spread on the chart, and it is one country against three. Poland scores 93: a strong need for rules, precision and plans, and a healthy suspicion of untested ideas. Ireland (35), India (40) and the US (46) are comfortable with ambiguity. American product teams treat it as freedom; Irish teams as room for ideas; Indian teams are famously good at what people there call “adjusting”, finding a way around a constraint that looked fixed. A Polish team that asks fifteen clarifying questions is not being difficult; it is doing the risk analysis the plan skipped. An Indian team that quietly adapts the plan to make it work is solving the problem, and also, sometimes, hiding it.

Indulgence: optimism and restraint

The US (68) and Ireland (65) score as indulgent cultures, optimistic and comfortable with enthusiasm. Poland (29) and India (26) score as restrained, cautious and wary of overclaiming. This shows up in tone and in status reports: American and Irish updates run green until they are suddenly red, while Polish and Indian ones run amber long after the risk is under control. Calibrate before you escalate.

Individualism: India’s team-first culture

The US (60) and Ireland (58) are individualist, Poland (47) sits in the middle, and India (24) is the most group-oriented of the four. Loyalty runs to the team and the manager, relationships shape decisions, and being singled out in public, for praise or for blame, can be uncomfortable. Recognition that lands well in Austin, one name on a slide, can land badly in India, where the team did the work and expects to be thanked as a team.

What all four share

Achievement and success (56 to 68) and long-term orientation (49 to 51) are close across all four; on long-term orientation the four scores sit within two points of each other. All four cultures are driven, competitive and pragmatic. People in every one of these offices want to win and want to do the work properly. That shared drive is the foundation, and it is why these teams work so well once the context problems are solved.

The twin trap

On paper, the US and Ireland look like twins: within twelve points of each other on every dimension. That is exactly what makes them easy to get wrong. Shared language hides real differences: Ireland is flatter than the US, more comfortable with ambiguity, and far more fond of understatement. An American who takes “it should be grand” at face value, or an Irish engineer who reads American enthusiasm as a commitment, is making the same mistake from opposite ends.

Setting context: what these numbers are, and are not

Country scores are useful and easy to misuse, so a few caveats before anyone builds a team on them.

• They describe the middle of a distribution, not your colleague. A national score is an average of survey answers. Treating it as a description of an individual is the ecological fallacy, and it turns a research tool into a stereotype.
• Some countries are many cultures. India’s scores average across more than a billion people, dozens of languages and very different regions and generations. A software engineer in one of its tech hubs may have more in common with a colleague in Kraków than with the national average.
• Your team is not the national average. Engineering has a culture of its own, and it is lower-context and more egalitarian than most of the countries it works in.
• The numbers move. Scores are periodically recalibrated with newer data. In its 2023 update, The Culture Factor cut the US individualism score from 91, the figure most textbooks still quote, to 60. Check the current scores before you quote old ones.
• Use them as hypotheses. When a team keeps clashing in the same place, the dimensions give you a vocabulary for why, so the conversation can be about the gap instead of about the people.

That last point is why I like the Culture Compass more than the country charts. It is a 42-question assessment of your own preferences. You choose your role, such as manager, colleague or negotiator, and up to three countries per report, and it shows the three dimensions where you differ most from each one. A manager working across all four of these cultures needs two runs. What matters on a distributed team is not any country’s score; it is the gap between a particular manager and the people they lead. An American leader with low uncertainty avoidance leading a team in Kraków, or an Irish leader with very low power distance leading a team in India, should know that before the first planning cycle, not after the third missed expectation.

Where AI changes the equation

A large language model is the most low-context colleague you will ever have. It knows nothing about your project that you did not write down. That makes AI an unexpected ally for distributed teams: the discipline it demands, explicit goals, written constraints, a clear definition of done, is exactly the discipline cross-cultural teams have always needed. I have argued before that context is the real work; teams that learn to write good context for agents write better context for each other.

Where it helps

• Intent, written down. A one-page statement of the goal, the constraints and what must not happen serves an agent, an engineer in Kraków, a team in India and a product manager in Austin equally well. High uncertainty-avoidance teams get the clarity they need; high-context teams get the background nobody thought to say out loud.
• Handoffs across the time zones. Austin is about seven hours behind Kraków, six behind Ireland and ten and a half to eleven and a half behind India, depending on the season. Austin and India barely overlap at all; Europe is the bridge, sharing working hours with both. AI summaries of each region’s threads, decisions and open questions make a three-hop, follow-the-sun handoff workable, and let the overlap hours be spent on conversation instead of status updates.
• Translating register, not just language. Before sending difficult feedback, ask a model how it is likely to land with a reader from the other culture. Use it to decode the message, not to sanitise it: what does “that’s grand” or a quiet “yes” actually mean here?
• Lowering the cost of dissent. Asynchronous design reviews, written pre-mortems and AI-collected risk lists give people a way to say “this will not work” without contradicting a senior leader in a live meeting. In the two high power-distance cultures, that is often the only way the warning arrives in time.

Where it hurts

• Flattening. Models default to an upbeat, hedged, American corporate register. Run a Polish engineer’s blunt and accurate assessment through one and the signal gets softened away. Run Irish understatement or Indian politeness through one and a real concern becomes indistinguishable from agreement.
• Stripping the context. High-context communication depends on relationship and subtext. A summary keeps the words and loses everything around them, which is exactly the part that mattered.
• False fluency. AI makes everyone’s English polished. Polished writing is not agreement, and it can hide the very disagreement you need to hear.
• Outsourcing the relationship. Summaries are not trust. The relational part of collaboration still happens in person and over time, and in group-oriented, high-context cultures it matters even more. Keep putting people on planes.

A short playbook

1. Have every manager on a multi-country team take the Culture Compass against the countries they work with, and discuss the gaps openly as a team.
2. Write intent for every significant piece of work. If an agent could not act on it, neither can a team ten hours away.
3. Never treat silence as agreement. Ask for risks in writing, before the meeting, by name, and make it safe to send them one to one.
4. Calibrate feedback in every direction. Teach Americans that Polish directness is not hostility, and teach everyone that Irish understatement and Indian politeness can carry a serious warning.
5. Recognise teams as well as individuals, especially where the culture is group-oriented.
6. Design handoffs around the bridge: let the European offices connect the American and Indian days, and use AI summaries to carry context across each hop.
7. Use AI to summarise, translate and decode, and never to rewrite someone’s judgement into something more comfortable.

The goal is not to make Poland, Ireland or India more like the United States, or the other way round. It is to make the implicit explicit, so every team can be itself and still build the same thing. AI makes writing context cheap. It does not make trust cheap, and leaders still have to earn that one the old way.

Sources

• The Culture Factor: Country Comparison Tool, India, Ireland, Poland and United States
• The Culture Factor: The Culture Compass
• Erin Meyer, The 8 keys for managing multicultural teams
• Erin Meyer on high- and low-context communication, including India
• Edward T. Hall, Beyond Culture (1976)
• Erin Meyer, The Culture Map (2014)