Rosalind Achterberg

Her observation is that most 'urgent' interruptions aren't actually urgent — they're synchronous by habit, not by necessity, because real-time chat and impromptu calls are the path of least resistance for whoever has the question. She replaces the default with written, async-first updates and a clear written escalation path for genuinely time-sensitive items, so the team doesn't need real-time presence to keep moving.

Instead of ad hoc check-ins, each person writes a short structured update at fixed points in the day — what's in progress, what's blocked, what's needed from others — and colleagues read updates on their own schedule rather than interrupting each other to ask. This shifts the coordination cost from the person doing deep work to the async document, which can be read anytime.

She acknowledges most people can't unilaterally change their whole team's culture, so her fallback is individual: block and clearly label deep work time on a shared calendar, publish your own async update at the start of it, and set a status that signals unavailability with a specific return time, rather than vaguely appearing 'busy.' A visible return time reduces the anxiety that drives people to interrupt anyway.

She is careful to define a narrow, explicit definition of what counts as urgent enough to interrupt a deep work block (a small named list, not "anything that feels important"), because an overly broad emergency exception quietly swallows the entire system.