1. LandOffer
  2. Blog
  3. All articles
  4. Create Team Working Agreements for Response Times, Focus Work, and Escalations

Onboarding & Teamwork

Create Team Working Agreements for Response Times, Focus Work, and Escalations

A completed fictional working agreement and two-week exception/revision record distinguish response from resolution, working hours from coverage, and a trial revision from proven reliability.

10 min readLandOffer team

Create team working agreements by defining which channel to use, when a response is expected, what that response must contain, and how urgent work reaches an available person. Separate acknowledgment from resolution. Protect focus time and off-hours, name coverage gaps, and trial the agreement before treating it as a dependable way of working.

A written target cannot create staffing or shared availability. This guide from LandOffer.ai follows a fictional team's completed agreement through a two-week trial. One request is acknowledged within its working-time window; another exposes an uncovered escalation route. The revision preserves that mixed result instead of claiming the team has established round-the-clock support.

Colleagues shaping a working agreement with explicit focus time and communication routes

Agree on the problem before choosing a response target

“Be responsive” can mean check every notification, answer by tomorrow, or solve the issue immediately. Ask what is actually failing: requests without owners, unclear deadlines, interrupted focus blocks, or urgent problems sent privately to someone asleep. These problems need different changes.

Discuss when people actually work, planned leave, dependencies, and existing organizational policies. A team can improve its own communication without promising a service level that it cannot staff. Production incidents, security concerns, or other formally managed events should follow the applicable organizational route rather than a new informal chat convention.

Atlassian's working-agreements play describes defining channel purposes, working preferences, and escalation decisions together, then revisiting agreements as conditions change. That supports a trial with explicit review. It does not supply a universal response window or establish that everyone can attend the same hours.

Include quieter and absent contributors before finalizing the agreement. Ask whether a target is feasible within their accepted working patterns. An agreement posted by a manager is not evidence that every named backup has accepted coverage.

Define the clock and the response

Specify whether a target uses elapsed time, the recipient's working hours, or staffed coverage hours. Explain what happens across evenings, weekends, local holidays and leave. Avoid “within one day” if one reader means a calendar day and another means eight working hours.

An acknowledgment should tell the requester what happens next. “I have this; I need the sample file; I will assess it by Tuesday at noon UTC” is different from “fixed.” State the resolution or assessment commitment separately from the response target.

GitLab's communication handbook states that its team members are not expected to respond outside planned working hours and describes recording current availability. Those are GitLab's practices. The general design question for this fictional team is how to make its own accepted availability and routing visible without treating a chat presence indicator as a duty roster.

Worked asset: a completed team agreement

The team, dates, events and records below are fictional. No real workplace messages or schedules were changed. Four people maintain an internal design-component library: Asha in Bengaluru, Leo in New York, Dara in London, and team lead Nora in London. The agreement covers ordinary team requests and time-sensitive internal blockers, not production incident response or customer support.

They accept a trial from November 2 through November 13, 2026, with a review on November 16. For this trial, Asha records 09:00–17:00 Asia/Kolkata, Leo 09:00–17:00 America/New_York, and Dara and Nora 09:00–17:00 Europe/London, Monday through Friday. Their recorded trial calendars have no local holiday exceptions; recorded leave still overrides these hours. These are this fictional team's agreed windows, not assumptions about people in those locations.

All shared deadlines include a date and time zone. Calendar displays must use the named location's rules for that date; the document does not assume offsets remain constant. Individuals can agree to other patterns and update the record before someone plans a dependency against them.

Their accepted version 1 agreement reads:

Area Completed agreement for the trial Limit or exception
Ordinary request Create a work item with context, requested action, one proposed owner, needed-by date/time zone, and consequence of delay; link it in team chat A proposed owner can redirect or decline; mention does not create acceptance
Ordinary response Proposed owner acknowledges or redirects within four of their scheduled working hours; acknowledgment names next step and a dated assessment point This is not a four-hour fix; off-hours, weekends and recorded leave do not count
Focus work Each person can reserve up to two consecutive working hours with ordinary notifications muted, displaying the block's end time Focus time still counts toward the four-working-hour response target; plan checks before/after the block
Time-sensitive blocker A blocker affecting at least two team members with a needed-by time earlier than the ordinary response window can use the staffed urgent route A preference for speed or a routine status question is not sufficient; explain impact and timing
Urgent coverage On trial weekdays, Dara is primary and Leo backup from 14:00–17:00 UTC; both have accepted these roles No team urgent coverage promised outside that window; leave requires an explicit coverage update
Urgent acknowledgment Post the work-item link and impact in the designated urgent thread; primary target is acknowledgment within 20 coverage minutes No resolution-time promise; if no acknowledgment after 10 minutes, contact listed backup; at 20 minutes, contact Nora
Off-hours Ordinary messages may be queued without expecting an immediate reply; schedule nonessential notifications where practical Do not privately chase an off-duty person; organization-managed incidents use their existing route
Decisions and review Nora resolves competing priorities; decision owner records the choice in the work item; team reviews exceptions November 16 Agreement changes need discussion with affected people, not silent edits to their obligations

Outside urgent coverage, a time-sensitive request remains visibly queued for the next staffed window. The requester informs the affected work's owner that a dependency is unconfirmed; nobody should schedule on an assumed immediate reply. If an organizational incident route applies, use it. Calling ordinary team work urgent does not create an incident responder.

The four-working-hour clock is also concrete. A request assigned to Dara at 15:30 London time on Friday, November 6 uses 1.5 scheduled hours before 17:00. The remaining 2.5 hours end at 11:30 London time on Monday, November 9. No response is expected over the weekend. If recorded leave removes Monday, the target must follow that changed calendar and the dependency may need rerouting.

Protect focus time without hiding dependencies

The GitLab guide to non-linear working illustrates how documented work can continue when colleagues are not simultaneously available. It also describes updating status around availability. This is an example of an asynchronous approach, not evidence that any particular schedule increases productivity or that every employer permits the same flexibility.

A useful focus signal says when a person plans to return, not why they need uninterrupted time. The team needs enough information to route work. A person can record unavailability without explaining private circumstances.

Focus protection also needs a genuine exception route. If the same person must continuously read urgent chat during a focus block, that block is not protected. Route qualifying requests to the accepted coverage person instead. If someone is covering the urgent route, they should not simultaneously promise an uninterrupted block that makes the target impossible.

Do not equate availability with permission to change another person's priorities. A response may expose a conflict that Nora must decide. Accepting an inquiry does not mean abandoning previously committed work without that decision.

A trial review keeping a missed escalation and an uncovered period visible

Worked asset: trial exceptions and the repaired agreement

The first trial record is an ordinary request. On Friday, November 6, at 15:30 London time, Leo asks Dara to assess whether a component's empty state needs an explanatory label. Dara is away from notifications during an agreed focus block until 16:30. She responds Monday at 10:00 London time, within the four-working-hour target, and commits to assessing the question by Tuesday at 12:00 UTC.

The acknowledgment does not settle the design. Leo records that his dependent review must wait for Dara's assessment. Nobody calls the weekend silence a failure. This one example gives the team a useful check on the wording, not proof that four hours is the right target for everyone.

The second record exposes a failure. On November 10 at 14:00 UTC, Evan, a London-based partner review coordinator using the published route, posts a qualifying blocker: two reviewers cannot finish their internal component review because they lack a required reference artifact, and the review cutoff is that afternoon. Dara is unexpectedly absent. Leo is on recorded leave, but the urgent roster still lists both of them without a coverage warning.

Evan contacts Leo at 14:10, sees the leave record, and contacts Nora at 14:20 as the agreement instructs. Nora, in another meeting, acknowledges at 14:35. The 20-minute target has been missed. The old roster did not provide real accepted coverage, and Evan followed the published steps, but the route did not deliver its stated target. The issue was missing coverage, not a duty for someone on leave to respond.

Nora records the gap and agrees with the review owner to postpone the cutoff rather than imply that the artifact is now available. The artifact remains missing at that update. Nora takes responsibility for correcting the coverage process; the team does not blame Dara for being absent or require Leo to respond during leave.

At the November 16 review, they complete this exception and revision record:

Trial evidence What it establishes Accepted version 2 revision
Dara acknowledges ordinary request Monday at 10:00 after Friday 15:30 assignment Target met using recorded working hours; resolution still pending at acknowledgment Keep the four-hour target for another trial; include the calculated acknowledgment deadline in dependent work items
Primary unexpectedly absent and backup on leave; roster unchanged A written roster can overstate coverage Before each urgent window, Nora confirms an available primary and backup with explicit acceptance; a static name is insufficient
Nora acknowledges at 14:35 after 14:00 request Original 20-minute target missed; escalation did not make the request resolved Mark the missed target honestly; Nora records the decision to replan the review cutoff and the still-missing dependency
No additional accepted backup is available for a later proposed window Capacity cannot be created by restating the target Display “no accepted urgent coverage”; queue ordinary work and ask its owner to replan time-sensitive dependencies
No new trial has occurred after the revision Version 2 is a proposed improvement, not a proven repair All four accept another trial through November 27, with review November 30; coverage and dependency exceptions remain review items

Nora's completed summary says:

We retain the ordinary working-hour response target and focus blocks. The November 10 urgent acknowledgment was late because the listed route had no effective coverage and the lead escalation was not immediately available. We have not treated that request as resolved by an acknowledgment or a postponed deadline.

Before future urgent windows, I will obtain explicit primary and backup acceptance and publish whether coverage exists. Nobody is assigned an off-hours or leave obligation by this agreement. If coverage cannot be staffed, we will show the gap and replan the dependency. We have agreed to trial this revision, not declared the escalation process reliable.

The revision corrects an actual-within-fiction process failure while retaining uncertainty. It does not demonstrate faster delivery, restore the missing artifact, or prove that the next urgent request will be acknowledged on time.

Make exceptions part of the next review

Review a small record of what happened: request type, receiving window, acknowledgment, unresolved dependency, and any interruption or missing coverage. Avoid turning the trial into surveillance of chat activity or a contest over who answers fastest.

Ask whether the agreement fits the team's real capacity. A repeatedly missed target may need a different target, more coverage, narrower scope, or changed delivery expectations. An exception caused by a confusing rule needs a rule repair, not a reminder to try harder.

For more practical teamwork guides, visit LandOffer.ai. Before adopting a response-time promise, test one ordinary request across off-hours and one urgent request with the primary unavailable. Write down who actually accepts the work and what the team will do when nobody can cover it.

Sources and Further Reading