Horizon-KnowledgeMake knowledge operational.
Build on the knowledge baseline created in Intelli-Core. IDP Studio connects sources, reviewed interpretations, saved cases and decision applications, keeping evidence and judgement visible.
AI enabled · Supported by your background knowledge
Coming soon, watch this space.
Coming soon, watch this space.
Start with the passage.
Review source passages and extraction proposals before they become part of the model. Keep the link back to what the source actually says.
Practical Wisdom, in practice
What makes Horizon-Knowledge different.
Feature 1 of 6
From documents to reviewed proposals
A source is the beginning, not an automatic conclusion.
Review grounding passages and extraction proposals before adopting them into the model. The source workspace helps connect vocabulary, rules and questions to the material from which they were interpreted.
This creates room for the domain expert to reject, qualify or investigate a proposal. Extraction is an aid to knowledge capture, not authority to decide what the organisation believes.
Review a policy clause, accept the supported rule and retain an ambiguous exception as an open question.
View full size ↗Feature 2 of 6
Reason over facts you have actually saved
Ask what follows from the model and the case.
Build a vocabulary and explicit rules, then save case observations. Run real IDP inference and optimisation, inspect explanations and retain the context of the run.
A saved result is evidence about that particular model and those inputs. It does not establish that the original source was interpreted correctly or that the model covers every real-world case.
Compare which configurations satisfy the rules for a saved customer requirement, then inspect why alternatives were excluded.
View full size ↗Feature 3 of 6
Compare interpretations without hiding disagreement
Keep judgement in the model-building process.
Logic review places pinned model and proposal snapshots beside their source passages. Record clause-linked questions, answers and competing interpretations rather than overwriting every disagreement.
The Interpretation lab combines versioned packages into profiles and runs them against the same saved case. Expert-review workflows support feedback and comparisons with explicit vocabulary mappings.
Run a conservative and an alternative interpretation over the same case and inspect exactly where their outcomes diverge.
View full size ↗Feature 4 of 6
Compose an application from reviewed knowledge
Let people use the knowledge without editing its rules.
Compose forms and charts into an application around the model. Guided build assists through staged work, including ambiguity review, formal checks, case comparison and expert validation.
AI-assisted changes are reviewable and use your own provider connection. Guidance remains available without an AI key; publishing an application does not by itself certify the model’s completeness or correctness.
Turn a reviewed product-configuration model into a guided form that explains the available choices.
View full size ↗Feature 5 of 6
Inspect the scope of a formal check
Keep “checked” precise.
Experimental local Lean checking can establish supported, scoped claims, including satisfaction for complete saved IDP candidates. Decision tables and trees expose classifications for supported conditions.
These results are not a universal proof of equivalence between representations, nor a proof of legal correctness. The value is in preserving what was checked, against which model, and within which supported scope.
Review a candidate that passed a local check while separately assessing whether the source rules were captured faithfully.
View full size ↗Feature 6 of 6
Share evidence, not just a conclusion
Make review possible outside the original session.
Open Knowledge File bundles can include selected saved runs, checks, decision views and comparisons. Offline review forms let experts comment without editing raw structured data.
Private case data and alternative interpretations have separate inclusion controls. Returned feedback still requires explicit import, keeping a reviewer’s comment distinct from an adopted change.
Send a specialist a bounded evidence package, collect feedback and decide what to incorporate into the next model revision.
View full size ↗Where your knowledge meets your judgement.
Broaden your Horizon.
Explore how Horizon-Knowledge can bring your proprietary knowledge into productive work.
Coming soon, watch this space.
AI and connected services may require your own provider accounts and separate usage charges. Feature availability depends on the product and build.
