Your quietest 1:1 might be a language problem, not an engagement one

A report who goes quiet in a 1:1 often gets read as disengaged. Sometimes that's right. But it's worth checking before landing there: if the meeting runs in that person's second language, a narrower explanation is on the table too. Language friction is one hypothesis worth checking when that silence shows up.

A DeepL-commissioned Censuswide survey of 600 business professionals aged 25+, fielded 7-18 May 2026 and split evenly between Belgium and the Netherlands, found 39-40% say they speak less during multilingual meetings, and a similar share (38-39%) say they avoid raising complex ideas because they struggle to put them into words ^source. That's a self-reported association at the group-meeting level, not a study of individual 1:1s, disengaged direct reports, or translation latency. Read one way, quiet is a motivation problem. Read the other way, it's a workflow problem wearing an engagement costume. Silence by itself can't tell you which.

Why this gets misread

A manager watching a quiet report has one signal to work with: what happens in the room. Silence during the hard part of the conversation is easy to file as checked out, since that's the visible pattern and there's rarely a competing signal in the moment to weigh it against. Language friction can produce that same silence for a different reason, and nothing in the room automatically flags which one it is.

A manager who suspects disengagement will typically probe harder, ask more open questions, maybe raise it at the next check-in. None of that addresses translation latency, if that's actually the cause. If language friction is the real driver, more probing in the same non-native language doesn't remove it; it just asks for more output in the language that was already the friction point.

The live call isn't the whole gap

It would be tempting to solve this with better live translation or captions in the meeting itself, and that helps. But it isn't the whole gap. A Wordly-commissioned Dimensional Research study of 205 US/UK enterprise event stakeholders (companies with 1,000+ employees, responsible for external events with 100+ attendees and/or internal meetings with 10+ attendees where more than 10% of attendees weren't native English speakers) found respondents rated a range of outputs as beneficial to event outcomes: 61% for high-quality real-time output during the call, and beyond it, 58% for full transcripts in all languages, 51% for meeting notes with action items, and 50% for marketing-ready summaries ^source. That's directional evidence that buyers of multilingual meeting tools value a durable record, not proof that such a record improves recurring 1:1 follow-through specifically; the sample was enterprise events, not distributed 1:1 pairs.

That distinction matters for a recurring 1:1 specifically. A single meeting is a live event. A 1:1 series is a relationship that depends on a record between sessions: what got decided last time, what's still open, what the report said three weeks ago that turned out to matter. Here's the hypothesis worth testing: if that record only exists in one person's second language, the report may have to keep re-translating their own history every time they prepare for the next session, in a way the manager doesn't.

A record only one side owns is a record that erodes

This resembles a familiar failure mode: any handoff where one side doesn't fully own the artifact tends to break down. A shared doc that only one person can edit fluently isn't really shared. Neither is a runbook written in a format only the on-call engineer understands; it doesn't survive the next incident. The hypothesis here is that a 1:1 record kept in a language only the manager is fully fluent in could have a similar problem: it might look like continuity while the follow-through actually depends on whoever owns the language, not whoever owns the work.

Over several sessions, that gap can plausibly compound. The manager remembers the thread because they wrote it and read it in their own language. The report has to reconstruct it in translation each time, and reconstruction is a likely place for detail to slip: the specific blocker, the exact commitment, the nuance that made last session's decision make sense. Treat this as a hypothesis, not a settled fact: what looks like the report forgetting may actually be the record failing to travel with them.

Where 1on1 puts the record

1on1 generates its session output, summaries and action items included, as multilingual output, not just a translated interface. English and Romanian are live today; German, French, Spanish, and Portuguese are scaffolded.

That's a design choice about where the record lives, not a claim about the rest of the category. The question worth asking is still the same one: after the meeting ends, does the record both people rely on to prepare for the next session live in a language the report is actually fluent in, or does it live in whichever language the tool happens to default to?

What to check on a distributed pair

If you manage a pair where the report's first language isn't the one the meeting runs in, the next 1:1 is a reasonable place to check something concrete: not whether the call went fine, but what language the record (not just the conversation) is actually in. Ask the report to pull up the summary from the last session before you do. If they're translating it in their head before they can use it, that's the friction, and it's worth testing with one pair before deciding it applies to the rest of the team. ^source