Modeling how knowledge degrades, fragmentation, custodian departure, lost context, and when to intervene.
Institutional knowledge does not fail all at once. It thins. Documentation falls out of date, the people who held the context depart, related artifacts are separated, and transformations accumulate without record until an asset that is still technically available is no longer actually usable.
We treat that decay as a modelable process with identifiable drivers, so that intervention can be scheduled before the loss becomes irreversible.
Four drivers recur. Provenance concentration, when the lineage of an asset depends on a single custodian or an isolated node, creates fragility. Structural isolation, where an asset has lost its relationships to the artifacts that made it interpretable, removes context. Temporal decay reflects the gradual loss of usability as an asset ages away from the environment that produced it. And departure hazard captures what is lost when the people bound to an asset leave.
Modeled together, these produce a calibrated estimate of the probability that an asset's knowledge becomes unusable, and importantly, an indication of which intervention would help most.
A lifecycle model earns its keep when it changes behavior. Our research connects degradation measurement to concrete preservation action: what should be documented, what should be re-linked to its context, what should be locked against alteration, and what should be moved to long-horizon retention.
Concentration and diversity across the lineage neighborhood.
Orphaned and disconnected assets that have lost interpretive context.
Usability modeled against a knowledge half-life.
Exposure created by custodian turnover and single points of knowledge.