Decision owner
Principal investigator
The named role remains accountable for the consequential decision.
Life Sciences & Pharmaceuticals
Designed for Research scientist
Let research scientists move from a biological question to a defensible evidence synthesis across papers, patents, protocols, and internal study reports, with a scientist reviewing provenance before conclusions are circulated.
Contract LIF-01 · primary aggregate ResearchQuestion · mode loading
Decision brief
This brief starts with the decision, evidence, and accountable owner. The guided workflow then shows how the recommendation was produced.
The evidence and control points that shape this recommendation.
Decision owner
Principal investigator
The named role remains accountable for the consequential decision.
Governed evidence
5 record types
Structured records and cited text ground the recommendation.
Workflow
5 controlled stages
Each transition records its input, output, status, and audit context.
Human controls
1 approval gate
Approval gates prevent the copilot from making consequential decisions.
A scientist asks whether a kinase target has selective inhibition data in solid tumour models; the copilot retrieves 240 publications, extracts 68 IC50 values with assay conditions, flags two studies whose selectivity claims conflict with the rest, grades the overall evidence as moderate, and the principal investigator approves the synthesis for the program review..
Let research scientists move from a biological question to a defensible evidence synthesis across papers, patents, protocols, and internal study reports, with a scientist reviewing provenance before conclusions are circulated.
Approved synthesis released to the program gives Principal investigator the evidence needed to begin the final governed business outcome, with the decision trail retained for review.
Step 1 of 5
| Field | Value |
|---|
| Field | Value |
|---|
The output appears here immediately. More technical evidence remains available below.
The demonstration keeps deterministic controls, AI assistance, and approval separate.
Source systems enforce identity, workflow order, policy filters, and record updates.
Retrieval and generation assemble relevant evidence into a cited recommendation or draft.
Principal investigator verifies material claims and approves the consequential action.
Run the stages in order before using the Copilot. Every stage explains what happens, why it matters, the evidence it uses, and the complete API request and response.
Automated stage
What happens: Research scientist performs frame the research question and scope using Target, modality, and inclusion criteria and produces ResearchQuestion with a defined corpus scope. The transition is deterministic and does not require autonomous model approval.
Why it matters: ResearchQuestion with a defined corpus scope gives Retrieval agent the evidence needed to begin Retrieve candidate literature and patents, with the decision trail retained for review.
Loading relevant sample records.
Ready to run.
POST /api/domain/workflow/frame-the-research-question-and-scope
Run this API to see its formatted outcome.
| Field | Value |
|---|
| Field | Value |
|---|
Automated stage
What happens: Retrieval agent performs retrieve candidate literature and patents using Question vector and metadata filters and produces Ranked LiteratureRecord and PatentFamilyRecord set. The transition is deterministic and does not require autonomous model approval.
Why it matters: Ranked LiteratureRecord and PatentFamilyRecord set gives Extraction agent the evidence needed to begin Extract quantitative findings, with the decision trail retained for review.
Loading relevant sample records.
Ready to run.
POST /api/domain/workflow/retrieve-candidate-literature-and-patents
Run this API to see its formatted outcome.
| Field | Value |
|---|
| Field | Value |
|---|
Automated stage
What happens: Extraction agent performs extract quantitative findings using Full-text sections and figure captions and produces AssayResultSummary records with units normalized. The transition is deterministic and does not require autonomous model approval.
Why it matters: AssayResultSummary records with units normalized gives Synthesis agent the evidence needed to begin Synthesize claims and flag contradictions, with the decision trail retained for review.
Loading relevant sample records.
Ready to run.
POST /api/domain/workflow/extract-quantitative-findings
Run this API to see its formatted outcome.
| Field | Value |
|---|
| Field | Value |
|---|
Automated stage
What happens: Synthesis agent performs synthesize claims and flag contradictions using Findings across studies and designs and produces EvidenceSynthesis with graded claims. The transition is deterministic and does not require autonomous model approval.
Why it matters: EvidenceSynthesis with graded claims gives Principal investigator the evidence needed to begin Scientific review and release, with the decision trail retained for review.
Loading relevant sample records.
Ready to run.
POST /api/domain/workflow/synthesize-claims-and-flag-contradictions
Run this API to see its formatted outcome.
| Field | Value |
|---|
| Field | Value |
|---|
Human approval required
What happens: Principal investigator performs scientific review and release using Synthesis with provenance and contradictions and produces Approved synthesis released to the program. A named human approver must review the evidence before the transition completes.
Why it matters: Approved synthesis released to the program gives Principal investigator the evidence needed to begin the final governed business outcome, with the decision trail retained for review.
Loading relevant sample records.
Ready to run.
POST /api/domain/workflow/scientific-review-and-release
Run this API to see its formatted outcome.
| Field | Value |
|---|
| Field | Value |
|---|
Run a stage to see its explanation, evidence, full request, and full response.
Ask a question and receive an answer grounded in retrieved ResearchQuestion evidence.
Select a starter question or enter your own.
Loaded live from /api/config, /api/domain/model, and
/api/domain/sample.
researchQuestions/tenantId + /researchProgramIdA scientist asks whether a kinase target has selective inhibition data in solid tumour models; the copilot retrieves 240 publications, extracts 68 IC50 values with assay conditions, flags two studies whose selectivity claims conflict with the rest, grades the overall evidence as moderate, and the principal investigator approves the synthesis for the program review.
| Container | Role | Hierarchical partition key | Document types |
|---|---|---|---|
researchQuestions |
Primary | /tenantId + /researchProgramId |
ResearchQuestion, EvidenceSynthesis |
scientificEvidenceCorpus |
Reference | /tenantId + /sourceDocumentId |
LiteratureRecord, AssayResultSummary, PatentFamilyRecord |
| Type | Container | Partition key | Identifier pattern | Embedding fields |
|---|---|---|---|---|
ResearchQuestion |
researchQuestions |
/tenantId + /researchProgramId |
question::{researchProgramId}::{questionId} |
questionVector |
LiteratureRecord |
scientificEvidenceCorpus |
/tenantId + /sourceDocumentId |
literature::{sourceDocumentId} |
abstractVector, titleVector |
AssayResultSummary |
scientificEvidenceCorpus |
/tenantId + /sourceDocumentId |
assay::{sourceDocumentId}::{resultIndex} |
None |
PatentFamilyRecord |
scientificEvidenceCorpus |
/tenantId + /sourceDocumentId |
patent::{sourceDocumentId} |
claimScopeVector |
EvidenceSynthesis |
researchQuestions |
/tenantId + /researchProgramId |
synthesis::{questionId}::{synthesisVersion} |
synthesisVector |
Knowledge-container query against scientificEvidenceCorpus (bounded fan-out across the /tenantId + /sourceDocumentId partitions named by the shortlist): gather the literature and assay evidence cited for the target, fanned out across the source document ids the programme's synthesis references.
scientificEvidenceCorpus
partition-key-fan-out
SELECT * FROM c WHERE c.tenantId = @tenantId AND c.sourceDocumentId IN (@citedSourceDocumentIds) AND c.type IN ('LiteratureRecord','AssayResultSummary') ORDER BY c.publicationDateUtc DESC
Knowledge-container query against scientificEvidenceCorpus (bounded fan-out across the /tenantId + /sourceDocumentId partitions named by the shortlist): screen extracted assay results against the potency threshold, over the bounded set of source documents retrieved for the question.
scientificEvidenceCorpus
partition-key-fan-out
SELECT * FROM c WHERE c.tenantId = @tenantId AND c.sourceDocumentId IN (@citedSourceDocumentIds) AND c.type = 'AssayResultSummary' AND c.normalizedValue <= @potencyThreshold ORDER BY c.measuredValue ASC
Knowledge-container query against scientificEvidenceCorpus (bounded fan-out across the /tenantId + /sourceDocumentId partitions named by the shortlist): surface live patent families with priority dates relevant to the programme's chemical space, over the patent family ids returned by the structure and claim search.
scientificEvidenceCorpus
partition-key-fan-out
SELECT * FROM c WHERE c.tenantId = @tenantId AND c.sourceDocumentId IN (@candidatePatentFamilyIds) AND c.type = 'PatentFamilyRecord' AND c.legalStatus = 'granted' AND c.priorityDateUtc > @priorityWindowUtc ORDER BY c.priorityDateUtc DESC
List syntheses for this programme whose claims conflict across studies and need scientist adjudication.
researchQuestions
single-partition
SELECT * FROM c WHERE c.tenantId = @tenantId AND c.researchProgramId = @researchProgramId AND c.type = 'EvidenceSynthesis' AND ARRAY_LENGTH(c.contradictionNotes) > 0 ORDER BY c.synthesisVersion DESC