Mnemos AI
Product

Continuity risk

Quantified, evidence-backed exposure scoring that surfaces what would break if specific people left.

Warning

Mnemos doesn't predict attrition. It quantifies the impact if attrition happens. Treat scores as exposure, not probability.

The risk score

Each entity (person, team, system, workflow) carries a continuity score from 0 to 100, where higher means more exposure. The score is a weighted sum of evidence-backed signals from the memory graph. No signal can fire without a citation.

SignalWeightWhen it fires
Single-person dependency0.35A critical workflow or system has exactly one named owner with no documented successor.
Undocumented workflow0.20A workflow is referenced in sessions or integrations but lacks an approved SOP.
Stale SOP0.15An approved SOP has dropped below the project's freshness threshold.
Tenure cliff0.15A long-tenured owner is within 90 days of a known departure or has flagged a planned exit.
Tribal contradiction0.10Multiple sources contradict each other on a workflow step; no resolution exists.
Vendor lock0.05A critical system has no documented in-house operator.

Dependency mapping

The platform precomputes the transitive closure of depends_on and owns edges per entity, so you can answer questions like "what breaks if Maya leaves?" without recomputing the graph at query time. Heatmaps highlight clusters where a small number of people own a disproportionate share of critical workflows.

Monetary exposure

Optional. If you configure workflow revenue impact (annual revenue touched, percentage at risk under disruption), Mnemos translates continuity scores into a dollar exposure number. We recommend calibrating these inputs with Finance once per quarter rather than per workflow.

Mitigations

For every elevated risk, the engine proposes a concrete mitigation: run a handoff interview, assign a successor, document a workflow, or schedule a freshness review. Mitigations are first-class tasks and feed back into the score as they're completed.