Plumbing call triage: burst pipe vs drain cleaning, a decision tree

Water everywhere. That's how the call usually starts, at midnight or on a Sunday morning — not a diagnosis, just a panicked homeowner describing the only thing they can see. Whoever answers has thirty seconds to decide: burst pipe, truck rolling now, or a slow drain that's finally gotten someone's attention after a week of ignoring it. Send a tech out at midnight for the drain and the crew is wrecked by sunrise for a job that could've waited. Tell the burst-pipe caller "we'll fit you in tomorrow" and there's a soaked subfloor and a customer calling the next plumber in the search results before anyone calls back. The gap isn't about how sharp the person on the phone is — it's whether there's a set order of questions instead of a fresh guess every time.

Quick answer: clear safety first — a gas smell or sewage backup is never triaged by severity, it's escalated immediately — then work through active water damage, whole-system versus single-fixture, whether this customer has called about the same problem before, and how much the issue is actually affecting the home today. If you're hearing symptoms instead of a specific fault, don't guess a tier; get someone on the phone who can ask follow-up questions. Every answer should carry a response window decided in advance, not on the call.

The rule that comes before triage

Two situations don't get triaged at all — they get escalated, full stop, before anyone asks a second question.

Gas smell. This is never a severity judgment call. The sequence is: get out of the house, call the gas company or 911, and only after that does anyone think about scheduling a plumber. A caller who mentions a gas smell should hear "leave the house now" before they hear anything about an appointment window. Service comes after safety, never instead of it.

Sewage backup. Raw sewage in a home is a health hazard, not a drainage inconvenience, and it doesn't get sorted by how bad it sounds on the phone. Any report of sewage coming up through a drain, toilet, or floor goes straight to a human for immediate handling — no attempt to estimate severity first, because the health risk is the same whether it's a trickle or a flood.

There's a third case that behaves differently from both of those: active flooding, water actively running, spraying, or pooling with the property at risk. Here the first instruction matters as much as the escalation — tell the caller to shut off the water at the main now, before anything else, because every minute the water keeps running is more damage to the floor, the walls, or whatever's underneath. Once that instruction is given, this still routes to a human immediately; a caller standing over an active leak doesn't need a diagnostic conversation, they need the water stopped and a plumber on the way.

None of this is a numbered spot on the tree below — it's a filter that runs first, before urgency gets ranked. A caller who mentions gas, sewage, or water spreading across a floor doesn't get sorted more carefully; they get a human immediately, and in the flooding case, one clear instruction before that handoff.

The decision tree: what comes after safety

Once a call clears the section above, the rest of triage is about ordering the remaining questions so a caller ends up in the right bucket without twenty minutes of back-and-forth.

1. Safety or health hazard? Gas and sewage escalate at the top of the article; nothing below applies to them.
2. Active water damage, and has the water been shut off? Water that's still running, spraying, or pooling — a burst line, a split pipe, a water heater that's let go — is an emergency regardless of what caused it. The status of the shutoff matters: a caller who's already killed the water at the main has bought some time; a caller who hasn't needs that instruction before anything else, then an immediate dispatch.
3. Whole system down, or one fixture? "No water pressure anywhere" and "the kitchen sink won't drain" aren't the same call, even if both sound urgent to the person reporting them. A whole-house failure — no pressure at any fixture, every toilet clogged, no hot water at all — moves faster through the queue than a single fixture acting up while the rest of the house is fine.
4. Has this customer called about this before? A slow drain reported once already and now worse, or a running toilet on a maintenance customer's third call this year, carries more weight than the identical complaint from someone calling in cold. Known history means the problem has a track record of getting worse, which is worth moving up the queue even when today's version sounds mild.
5. What's actually not working today? Separate "completely unusable" from "annoying but functional." A toilet that won't flush at all is a different call than one that runs an extra thirty seconds after each flush. Functional impact today, not the customer's frustration level, is what should move the needle.
6. Symptoms, or a specific problem? "The kitchen faucet is leaking" names a fixture and a fault, so it can be ranked against everything above it. "The water looks a little off" names neither, and ranking it means ranking a guess. Calls in that second shape need one more conversation before they can be sorted at all — someone who keeps asking until the fault has a name.

Some owners weigh customer history above single-fixture-versus-whole-system, and that's a defensible call to make differently. What can't move is the sequence itself, once it's set.

A starting triage matrix

Set your own hours against each row, then leave them alone — nobody should be negotiating urgency from scratch while a caller describes water spreading across a floor.

Where triage goes wrong

A plumber gets pulled out at 3 AM for a toilet that's been running since dinner — not overflowing, just loud enough that the homeowner couldn't sleep next to it. He's back in bed by four, except three nights like that in a row and he's not sharp on the fourth, which is exactly when a real emergency comes in. That's what over-triage costs: not the one wasted dispatch, but a crew too worn down to catch the call that matters.

Under-triage shows up on the other side of the same week. A tenant calls about water "spraying a little" under a bathroom sink, sounds calm enough on the phone, and gets penciled in for Thursday. By Wednesday night the cabinet's swollen and there's a stain spreading across the ceiling below — a twenty-minute fix on day one, now a claim with an adjuster attached. Nobody made a reckless call; they just never asked whether "spraying" meant a drip or a jet.

Assuming instead of asking connects both stories. "My plumbing isn't working" could mean a slow drain or zero pressure at every fixture — two calls that shouldn't share a queue position, but will if nobody asks the follow-up question. "Is it just the one sink, or everywhere?" would have sorted both calls correctly before either turned into a story worth telling.

The fix is the same written sequence, applied the same way by whoever answers — dispatcher, answering service, or configured AI — instead of a fresh judgment made under pressure on every call. When something genuinely doesn't fit cleanly into a tier, treat that ambiguity as the reason to loop in a person, not as permission to pick a category and move on.

How this works with an AI front office

Picture the 12:30 AM version of this call landing on Patchment (that's us — worth knowing we have a stake in how this paragraph reads). A caller says water's spraying under a sink, and the first thing they hear back isn't a diagnostic question — it's the instruction to shut the water off at the main, right now, because that's the shop's own rule for active flooding. The call then routes to a human before anything gets booked. Nothing there was invented on the spot: every threshold underneath it — what earns a same-day slot, what gets scheduled out, what requires a person before a slot is confirmed — was set by the owner ahead of time. Two things hold regardless of the call: Patchment never quotes a price over the phone, and anything safety-adjacent — gas, sewage, active flooding — reaches a human immediately instead of getting resolved on its own. It runs the shop's playbook; it doesn't write one.

This triage sequence assumes someone or something already picked up — for the separate question of why plumbing calls go unanswered in the first place, see why plumbers miss emergency calls. For how AI call answering handles intake and routing across the rest of a shop's calls, our guide to AI call answering for field service walks through that. Patchment's plumbing page covers the trade-specific configuration.

The real test of a triage system isn't the tiers — it's whether "shut off the water at the main" gets said before anything else, every time a caller describes water that's still moving. Get that instruction out first, consistently, and the matrix above did its job; skip it once, and the matrix never mattered — the subfloor was already ruined before anyone reached for it.

Book a demo to watch a shut-off instruction go out before a single tier gets assigned.