Team Decision Scenario · Pressure Signal: Diffusion
Who Decides When No One Answers?
It’s Friday at 4:00. The platform is down on payday, leadership is unreachable, and the system that broke it can’t tell you what it did. Now your team has to decide who acts — before the clock and the weekend run out.
Quick Answer
What should a team do when a system goes down, and no one with authority is reachable?
Organize around whoever can work the problem, act on whatever documented authority exists, and escalate—and document everything else—don’t let the group freeze waiting for a sign-off that isn’t coming. But the real fix is upstream: decide in advance who holds authority when leaders are unreachable, and ensure that the people who understand critical systems are available during an incident. A team can survive the night by improvising. It shouldn’t have to.
The Pressure Signal: Diffusion
In a leaderless group under pressure, responsibility diffuses. Everyone assumes someone else will make the call — so no one does. The bigger the group and the higher the stakes, the stronger the pull. It isn’t a lack of competence. It’s a lack of a designated decider.
The Situation
Hollensby Financial Systems runs the platform that moves payroll for hundreds of client businesses. At 4:00 on a Friday — payday for much of the client base — the platform goes dark. A top-ten client is already on the phone: their employees can’t be paid.
The CEO is on an international flight, unreachable. The Chief Revenue Officer and CHRO are at the off-site top-performer trip, off the grid. And the root cause isn’t a person you can call: an autonomous agent with production access — deployed weeks ago for routine maintenance — executed a change no one authorized in the moment. It did exactly what it was scoped to do. There’s no engineer who “ran the wrong command” to interview. There’s only behavior to reconstruct from logs.
The people who could read those logs — the two engineers who built and scoped the agent — are remote and logged off for the weekend, notifications silenced. In the room: an on-call ops lead, a support lead fielding the client, a junior engineer, and the VP of Engineering, who has a hard stop at 6:00 to pick up his kids.
“A good team doesn’t need to be told to jump on an outage. Ours would.”
True — and not the point. Jumping in isn’t the same as having the authority to make the call that fixes it. “We’d figure out who stays” is easy to say at 2:00 and hard at 5:15, when the one person who understands the system has to leave at 6:00, and the people who could replace him are offline. The question isn’t whether your team cares. It’s whether anyone in the room is actually allowed to decide — and whether enough of the right people are still reachable to act.
Who’s in the Room — and What They Can Authorize
Run this as a group. Have each person read their seat aloud before the team decides.
Mara Quinn — On-call Ops Lead. Has been running incident response by default. Can follow runbooks — but a production rollback that risks data integrity is above her documented authority.
Soren Adler — VP of Engineering. Has the authority and the knowledge to read what the agent did. Has a hard stop at 6:00 to pick up his kids.
Devon Marsh — Junior Engineer. Capable hands and willing to stay all night — but doesn’t know the agent’s deployment and can’t authorize a production change.
Hugh Daley — Support Lead. Owns the client relationship in the moment. Can’t authorize technical action, but has to tell a top-ten client something — and what he promises commits the company.
Not reachable: the CEO (in flight), the CRO and CHRO (offsite), and the two engineers who built the agent (remote, logged off, notifications silenced).
The Decisions Your Team Has to Make
Work them in order. There is no answer key — the discussion is the point.
4:00 — Who’s in charge?
The outage is confirmed, and leadership is unreachable. Mara has been running point by default — but the call that might fix it (a risky rollback) is above her authority. Does someone formally own this incident? Who? And how long do you spend trying to reach leadership before you act without them?
5:15 — The bench thins
The fix isn’t closed. Soren leaves at 6:00. The two engineers who understand the agent are offline and not answering. Do you ask Soren to stay — and can anyone make him? Do you act with the people physically present, who may not fully understand what the agent did? How long do you chase someone who’s already gone for the weekend?
6:00 — The handoff
If Soren goes, how does he hand off something that doesn’t hand off in ten minutes — how to reconstruct what a non-human actor did — to a team that may not include anyone who can read it? What gets written down? What does the client get told in the meantime, and on whose authority?
What Good Looks Like (It’s a Recognition, Not an Answer)
A capable team gets through the night the way capable teams always have: someone takes ownership — even informally — the group acts on whatever authority is actually documented, and escalates and records everything that isn’t. They protect the recovery with a clear, written handoff so knowledge doesn’t walk out the door at 6:00.
But notice what just happened: every one of those moves was improvised. None of it was decided in advance. The real failure wasn’t a person leaving for their kids or logging off at 5:00 — everyone did something reasonable. It’s that the organization never decided who holds authority when leaders are away, never made “reachable during an incident” mean anything, and never planned for an actor that can’t be interviewed. The right call is to survive the night and name those gaps before the next Friday.
The Scenario Is Just the Vehicle
A Friday outage caused by an AI agent, with no reachable authority and a thinning team.
What It’s Really About
Whether your team has a designated decider — and whether the right people are reachable — when the moment actually comes.
The real discussion is not about a payment platform or an AI agent. It is about who acts when authority isn’t in the room, and what your continuity plan quietly assumes about people who may not be there.
How to Run This With Your Team
Set aside 20–30 minutes. Read the situation aloud, assign the seats (or let people pick), and work the three decision points in order. Don’t rush to a fix — the value is in watching where the group hesitates over who’s allowed to decide. Close by writing down the gaps the exercise exposed: Who is our designated incident authority when leaders are unreachable? Who must be reachable, and who ensures coverage? Does our continuity plan assume people will be available who might not be?
Then take those questions to the next level. This scenario is the team-sized version of an executive exercise we built for exactly this gap.
Pressure-Test It at the Top
Your team just felt the open chair. Does your leadership know they have one?
The Executive Decision Lab™ runs this same crisis at the C-suite level — a 90-minute facilitated session that surfaces whether authority actually transfers when leaders are unreachable, and whether your business continuity plan survives an actor it can’t interview.
Frequently Asked Questions
Who is in charge during an outage when leadership can’t be reached?
Whoever your organization designated in advance — ideally a named incident commander with pre-defined authority to act. If no one was designated, responsibility tends to diffuse and the group waits. The fix is to decide and document this before an incident, not during one.
Why is an AI agent harder to recover from than a human mistake?
A human who made a change can tell you what they did. An autonomous agent can’t be interviewed — the team has to reconstruct its actions from logs, and often only a few people understand the deployment well enough to do it. If those people are unavailable, recovery slows dramatically.
How do you run this as a team exercise?
Take 20–30 minutes, assign the seats, and work the three decision points as a group. There’s no answer key — close by writing down the authority and availability gaps it exposes, and feed those into your incident response and business continuity plans.
Want more scenarios that build decision-ready teams? Browse the free Scenario Library or see the full Xcelus approach.
© 2005–2026 Xcelus LLC. All rights reserved.