The single point of failure, human version
In IT, a component without redundancy is a known design fault. In an organisation the same fault happens in silence: a transfer, a departure, a reorganisation, and a task four people could do is now known by one. No dashboard shows it, because no data carries it.
How to make it visible
- Attach each piece of knowledge to the tasks it serves.
- Assess who masters it, from 0 to 5.
- Let availability compute itself from the real number of people assessed at a sufficient level. It is never typed in by hand.
The list of single holders then comes out by itself, ranked by criticality. On the demonstration it counts 33 for 84 tracked pieces of knowledge.
What to do next
For each single-holder piece of knowledge, the decision board proposes a successor: the closest person in skills, in the same structure when possible. An action plan sets the pair and the deadline; the risk only comes down once the successor has been assessed at the required level.
Why a declaration is not enough
A manager can declare a risk "covered". He cannot declare that someone knows: the person must have been assessed. That constraint is what turns the map from a negotiation into a measurement.
The real network
The internal social network — who really consults whom — reveals the people the organisation truly depends on. They are rarely where the org chart expects them.
Critical knowledge with no document at all is the most exposed case. Once your systems are connected — SharePoint, document management, Notion, intranet — the list of red-zone knowledge with no reference document shows up by itself.