Vagus is institutional memory for organizations, so that anyone can ask why a decision was made and get who decided, when, and on which sources.
Append source-connected claims, episodes, lessons and measured failures as distinct records.
Attributed original or revised commitment.
Append source-connected claims, episodes, lessons and measured failures as distinct records.
Source, holder, observation time and record identity.
Supersession, contest and retraction retain the earlier record.
Revision, contest and retraction preserve the earlier source history.
Resolve validity, knowledge time, scope and conflicting authority for the requested question.
Scope, validity time and knowledge cutoff.
Claims and operations with source ancestry.
Resolve validity, knowledge time, scope and conflicting authority for the requested question.
Event validity and recorded-at cutoff.
Authority ancestry and unresolved conflicting claims.
Both the validity question and the knowledge cutoff govern which records apply.
Find relevant evidence and assemble a bounded source-bearing answer.
Applicable, contested or unavailable evidence.
Rebuilt index/packs and retained failure context.
Find relevant evidence and assemble a bounded source-bearing answer.
Canonical records remain the historical basis.
Item/byte bounds, source pointers and abstention reasons.
The answer stays within the selected scope and may abstain when evidence is inadequate.
Integrate eligible change, retain standing causes and distinguish blindness from resolution.
Eligible changed source, freshness and provenance.
Integrate eligible change, retain standing causes and distinguish blindness from resolution.
Charge, surprise, decay and evidence trust.
New, standing, stale and cleared states.
Stale or missing sensing does not clear a standing issue.
Compare a stored estimate with a later identified outcome and retain uncertainty.
Defined target and context known before execution.
Independent outcome under the stored task identity.
Compare a stored estimate with a later identified outcome and retain uncertainty.
Defined target, task identity and resource estimate.
Independent outcome, unknown endings and eligible-choice advice.
Outcome advice remains separate from permission to act.
Rebuild derived context and retain useful failures without replacing source history.
Canonical records and their integrity state.
Attributed retrieval use, not proof of future utility.
Rebuild derived context and retain useful failures without replacing source history.
Duplicate/conflict annotations and derived packs.
Measured harms and reasons for caution or release.
Derived indexes can be rebuilt while canonical history remains intact.
Attributed original or revised commitment.
Append source-connected claims, episodes, lessons and measured failures as distinct records.
Source, holder, observation time and record identity.
Supersession, contest and retraction retain the earlier record.
Revision, contest and retraction preserve the earlier source history.
Scope, validity time and knowledge cutoff.
Claims and operations with source ancestry.
Resolve validity, knowledge time, scope and conflicting authority for the requested question.
Event validity and recorded-at cutoff.
Authority ancestry and unresolved conflicting claims.
Both the validity question and the knowledge cutoff govern which records apply.
Applicable, contested or unavailable evidence.
Rebuilt index/packs and retained failure context.
Find relevant evidence and assemble a bounded source-bearing answer.
Canonical records remain the historical basis.
Item/byte bounds, source pointers and abstention reasons.
The answer stays within the selected scope and may abstain when evidence is inadequate.
Eligible changed source, freshness and provenance.
Integrate eligible change, retain standing causes and distinguish blindness from resolution.
Charge, surprise, decay and evidence trust.
New, standing, stale and cleared states.
Stale or missing sensing does not clear a standing issue.
Defined target and context known before execution.
Independent outcome under the stored task identity.
Compare a stored estimate with a later identified outcome and retain uncertainty.
Defined target, task identity and resource estimate.
Independent outcome, unknown endings and eligible-choice advice.
Outcome advice remains separate from permission to act.
Canonical records and their integrity state.
Attributed retrieval use, not proof of future utility.
Rebuild derived context and retain useful failures without replacing source history.
Duplicate/conflict annotations and derived packs.
Measured harms and reasons for caution or release.
Derived indexes can be rebuilt while canonical history remains intact.
An illustrative process for source-grounded recall, attention and outcome comparison.
Additional mechanisms are described in the technical reference.
Dated records support recall of both earlier and current knowledge. Changes accumulate into attention, while later outcomes are compared with retained estimates.
Keep episodes, lessons, claims and failure memory with their provenance and holder.
Integrate signals over time, with decay, refractory behavior and explicit unknown states.
Compare a retained estimate with a later identified outcome. Keep uncertainty and useful failures available for the next decision.
An organization should not have to learn the same lesson twice.
Used for Naevum’s memory and work recordsDiscuss institutional intelligenceWhat the research says
More than half of workers say asking a colleague is often the only way to get the information they need. When that colleague leaves, the answer leaves too: across 110 studies, higher staff turnover goes with lower organizational performance, a modest link. In one benchmark, commercial chat assistants lost 30% accuracy recalling information across sustained interactions, so a conversation log is a weak substitute. Vagus records each decision with its date, its authority and its sources, so the answer stays when the person goes.123
A delivery date moves. The new date is current, but the old commitment still explains yesterday’s plan.
A temporal claim is resolved within its validity interval. A historical question can recover the claim that applied at that time.
A replacement claim changes the current answer. Source pointers and contested claims preserve the reason for that change.
A large organization produces more signals than anyone can examine individually.
Signals charge a subject’s attention potential. Decay, thresholds and refractory periods shape when it becomes worth examining.
A quiet source and an unobserved source are different conditions. Missing observability can itself contribute an attention signal.
Measured harms create decaying failure contexts. Hysteresis can preserve appropriate caution as conditions change; a warning cannot grant a new capability.
The COO asks; memory answers with the decision, its date, its authority and the sources of that time. Illustrative.
Illustrative situations, connected to the mechanisms that make them possible.
Institutions
A policy changes, but earlier claims and the reasons for past decisions still matter.
Engineering organizations
Signals arrive from many projects, workers and resource windows.
Operational teams
Different attempts consume different resources and produce different outcomes.
Vagus connects memory and attention to estimates of task outcomes, elapsed time and resource use.
Predictions of task outcomes, elapsed time and resource use enter the record before the outcome arrives. Later observations test the original estimate; they cannot silently replace it with a prediction written after the fact.
Hierarchical models can share evidence across a portfolio, organization, project and worker while retaining the uncertainty of a small sample. Worker selection can explore eligible options; an empty evidence base cannot support a confident winner.
The memory index can be rebuilt from source-connected claims, lessons, failures and episodes. A derived lesson remains distinguishable from a binding instruction. Useful evaluation asks whether people recover the right evidence and make better-supported decisions.
Document retrieval is one part. Vagus adds temporal claims, operating signals, attention and outcome context so that a source can influence ongoing work while keeping its history.
Choose a bounded set of sources and recurring decisions. Establish access boundaries, a fixed evaluation question set and the workflow where the resulting evidence will be used.
The technical choices that give Vagus its particular behavior.
Vagus connects temporal knowledge to attention and operational decisions. Claims retain their source and validity; signals accumulate over time; outcome estimates keep their uncertainty.
At the earlier date. A temporal claim is resolved within its validity interval. A historical question can recover the claim that applied at that time.
With today’s evidence. A replacement claim changes the current answer. Source pointers and contested claims preserve the reason for that change.
Accumulate relevance. Signals charge a subject’s attention potential. Decay, thresholds and refractory periods shape when it becomes worth examining.
Notice missing visibility. A quiet source and an unobserved source are different conditions. Missing observability can itself contribute an attention signal.
Remember the failure. Measured harms create decaying failure contexts. Hysteresis can preserve appropriate caution as conditions change; a warning cannot grant a new capability.
An illustrative institutional question that needs both history and controlled model context.
Follow the storyWhich past decision does your team argue about again because nobody can find why it was made?
Discuss institutional intelligence