Escalation Path Integrity Check
Your system escalates. Find out whether it reaches somebody who can actually decide — and whether their answer changes anything.
Most escalations reach somewhere. Whether that somewhere holds the authority to say no, and whether the no is enforced by anything, are two different questions, and almost nobody has asked the second.
instant download · .xlsx · yours to keep
The problem
It escalated. Nothing happened.
The refund was above the ceiling, so the system escalated. The message arrived in a channel four people are in. Two assumed the others had it. The one who read it properly does not have the authority to approve a refund that size and would have had to find somebody who does. By then the automatic timeout had passed and the action went ahead anyway. Every part of that worked exactly as designed.
Does it arrive at somebody who can act, and can their answer change what happens. A path takes the worse of the two.
Two paths in the worked example are identical except for whether the recipient can decide it themselves. One reaches someone. One reaches a queue.
The busiest path in the example works and the rarest works too. How often a path is used is reported and orders nothing.
What's inside
One spreadsheet, five answers per path.
- Where does the escalation land — a named rota, a shared queue, or somewhere unwatched?
- Can whoever is there decide this, or do they have to find somebody else?
- And outside working hours?
- If they say no, what makes that stick — or does the calling code simply choose to honour it?
- What happens if nobody answers at all?
The standard
Authority, not attention.
It is better than nowhere, so it sits above REACHES NOBODY rather than beside it. But nobody there owns this particular escalation, which is the property that fails at 2am.
An escalation that arrives at somebody who then has to go and find a director is a notification with extra steps. The question is whether the recipient can say yes or no themselves.
If the calling code reads the human's answer and chooses to honour it, that is a convention. The same custody question applies to a person's refusal as to a gate's verdict.
How it works
Each half takes its weakest answer.
Three answers make up the way out and two make up the way back. Each half reads as its weakest answer, and the path reads as the worse of the two halves. Both are always shown, because the most common shape in practice is a path that is flawless on the way out and has no way back at all — and a single blended score would hide exactly that.
What you'll see
Change one answer and watch a path stop reaching anyone.
The busiest path in this example is a working one and the rarest is too. How often a path is used orders nothing.
This is the live engine. The full .xlsx, with every path you have and both halves scored A Start Here tab that explains what each answer means A dashboard that names the one path to fix first
Get the kit — $59Who it's for
For the handoff nobody has watched fire.
- Something automated in your business escalates to a human when it is unsure.
- You could not say, today, whether the person it reaches can approve the thing being escalated.
- You have never watched what happens when nobody answers.
- You are building the human-review branch of an automated decision and want the contract settled.
- Grading whether your human-review controls are well designed. This asks a narrower question.
- Deciding whether the person answers correctly. It never looks at that.
- Call handling, or when a voice bot should hand a conversation to an agent. Different domain.
Scope: an engineering instrument. It grades an escalation path, never a person, asserts no regulatory position, and is not legal advice.
Common questions
Direct answers, before you buy.
Get it
An hour with your own paths.
- One .xlsx, three tabs, live formulas, worked example pre-filled.
- Both halves scored separately on every path.
- Offline. No install, no login, nothing to connect.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai