Content sources
Whatever the organization already reads: internal documents, published articles, standards, code, and research.
Design Pattern · September 2026
Lectio
The reference implementation of the closed loop: make a knowledge-sharing request explicit, observable, and answerable, so delivery, consumption, comprehension, application, and outcome stay separate claims with separate evidence.
Under the pillar Human knowledge as a closed loop
The pattern
Lectio makes one act explicit: a person asks another person to read something for a reason. Today that act is a link in a message, and it leaves no trace anyone can act on. Lectio turns it into a first-class object with a requester, a reader, a reason, a passage, and an expected response.
Once the request is an object, everything downstream becomes observable without surveillance theatre. The reader sees why the passage was sent and what response is wanted. The requester sees whether it was read and what came back. Neither has to guess.
Lectio runs inside netspective.com rather than as a separate learning system, because knowledge-sharing belongs where the work happens. A disconnected LMS reintroduces the exact gap the pattern exists to close.
Responsibilities
The pattern separates concerns so each one can be replaced without breaking the loop.
Ingestion and normalization
Bring the source in, keep its provenance, and reduce it to passages small enough to be sent for one purpose.
Precision metadata
What the passage is about, who it is for, what task it supports, and what a competent response to it looks like.
Identity and learner state
Who the reader is, their context, and what the system already believes about their knowledge.
The reading experience
The passage, the reason it arrived, the requested response, and the ability to answer in place.
Evidence and analytics
Delivery, open, read, response, and application recorded as separate events, then rolled up into learner state.
Three phases
Phase one proves consumption. Instrument explicit requests and separate delivered, opened, and read. Most organizations cannot do this today, and it is worth doing on its own because it kills the fiction of the completion report.
Phase two proves application. Tie the passage to a defined response and a work artifact, with a rubric that a person or a model can apply consistently. This is the phase that makes the data worth acting on.
Phase three closes the loop. Maintain learner state, recommend the next best item from evidence, and detect knowledge gaps early enough to do something about them.
Read, understood, applied
Keep the three claims apart. Read is an observation of attention. Understood is a response judged against a rubric. Applied is an artifact produced in the course of real work. Collapsing them is what makes conventional training reporting useless.
Reported separately, they are immediately useful to a manager: the passage went out and nobody read it is a distribution problem; everyone read it and nobody could answer is a content problem; everyone answered and nothing changed in the work is a systems problem.
How it gets built
Lectio is built AI Native: one developer owns the complete deliverable, starting from an executable specification and using an engineering harness for implementation, testing, investigation, refactoring, documentation, and recurring audits.
That matters to the research, not just the schedule. The specification is the artifact under review, so the loop being described and the system implementing it are held to the same standard of evidence.
Audit questions
Every review asks the same set. Can the system tell delivered from read? Is there a response tied to the passage? Is there an artifact tied to the response? Did a metric the business already tracks move? Did learner state change as a result, and did the next recommendation reflect it?
Stop or continue is decided on those answers. If a phase cannot produce its evidence, the next phase does not start, because a recommendation built on an unproven claim is worse than no recommendation.