01
Decisions disappear
Endpoint logic, mapping exceptions, imputation choices, and review resolutions are often recorded once, scattered, or reconstructed later.
Clinical Trials + Clinical AI
Persistent memory for clinical trials.
Ix connects clinical documents, code, assumptions, mappings, recommendations, and decision history into a persistent knowledge graph so teams can explain where outputs came from, what changed, and which reasoning should be reused.
Clinical documents
Code + derivations
Ix graph memory
Decision history
AI outputs
The Problem
Clinical trials preserve data. They lose reasoning.
Clinical trials generate more than datasets. They generate decisions: endpoint definitions, missing-data assumptions, mapping choices, protocol amendments, imputation logic, review comments, and analysis tradeoffs.
The final tables, listings, figures, and submissions may survive. But the reasoning behind them often lives across spreadsheets, SAS comments, emails, slide decks, Slack threads, and the memory of people who may not be on the next study.
02
AI sees fragments
LLMs can retrieve text, but they do not automatically understand how documents, code, data, assumptions, and decisions relate.
03
Review becomes reconstruction
Teams spend time re-explaining why a result exists instead of traversing a durable record of how it was produced.
The Solution
Ix makes clinical decisions first-class.
Ix builds a persistent knowledge graph across clinical documents, source code, data workflows, assumptions, recommendations, and decision history.
Instead of forcing teams and AI systems to reconstruct context from disconnected files, Ix gives them a structured map of what happened, why it happened, what changed, and what downstream outputs were affected.
01
Clinical sources
02
Ix Knowledge Graph
03
AI / analyst workflow
04
Traceable explanation
Core Capabilities
What Ix adds to clinical trial workflows
Recommendation Traceability
Trace clinical AI outputs back to source documents, assumptions, and reasoning steps.
Decision Lineage
Represent endpoint definitions, mapping rules, missing-data assumptions, and analysis choices as connected, versioned objects.
Cross-Study Memory
Reuse prior protocol patterns, mappings, edge cases, and decision history across studies and therapeutic areas.
Change Awareness
Show what changed across protocols, source documents, code, assumptions, or constraints, and what those changes affect.
Source-Backed AI
Give AI systems provenance-backed context instead of isolated prompts, scattered files, or one-off retrieval.
Defensible Outputs
Help clinical, regulatory, and enterprise teams understand why a recommendation, analysis, or workflow decision was made.
Where Ix Fits
From scattered clinical artifacts to traceable AI output.
01
Clinical Sources
Protocols, SAPs, CRFs, reports, literature, emails, ELNs, code
02
Ix Knowledge Graph
Persistent memory, structured context, source paths, decision lineage, version history
03
AI or Analyst Workflow
Recommendation, analysis, review, protocol design, trial planning
04
Traceable Explanation
Evidence, assumptions, reasoning path, version history, reusable context
Clinical Use Cases
Built for the places where clinical reasoning gets lost.
Clinical AI Recommendations
Trace every recommendation back to source material, assumptions, constraints, and prior decision context.
Protocol and Study Design
Reuse previous protocol patterns, endpoint logic, inclusion criteria, constraints, and amendment history.
Biostatistics Workflows
Preserve reasoning around endpoints, imputation choices, missingness assumptions, sensitivity analyses, and sample-size logic.
Clinical Data Mapping
Capture CRF-to-CDISC mapping decisions, exceptions, reconciliation logic, and downstream impact.
Regulatory Review Support
Produce clearer, repeatable explanations that show why decisions were made and what evidence supported them.
Enterprise Knowledge Continuity
Preserve institutional knowledge across vendors, teams, studies, clients, and sessions.
Knowledge Graph Model
AI needs a map, not another folder.
Ix uses a scene-graph-inspired model to represent entities, relationships, hierarchy, and time. In clinical trials, this means protocols, endpoints, visits, derivations, source documents, code, assumptions, decisions, and outputs can be connected as structured objects instead of stored as disconnected files.
Entities
Typed objects such as protocol, endpoint, visit, derivation, source document, code module, recommendation, or decision.
Relationships
Explicit links such as derived_from, mapped_to, amended_by, justified_by, depends_on, or changed_by.
Hierarchy
Studies contain sites, protocols contain endpoints, endpoints depend on derivations, and outputs depend on decisions.
Time
Every node and relationship can carry version history so the system remembers what changed and when.
Open Source Foundation
Built on open-source codebase intelligence.
Ix began as a local-first codebase intelligence system that maps repositories into persistent architectural graphs. Developers use Ix to understand structure, trace impact, explain behavior, and give AI agents durable context across sessions.
The clinical trials page extends that same graph foundation beyond code, connecting clinical documents, source evidence, data workflows, decisions, and AI outputs into one traceable system.
Open-source foundation
Local-first code mapping
Persistent graph memory
AI-agent compatible context
Impact tracing across code and documents
Designed for source-backed reasoning
Pilot
Start with one clinical workflow.
Ix can begin with a focused pilot around one high-value workflow: recommendation traceability, protocol reasoning, clinical document memory, code-to-decision lineage, or cross-study context reuse.
Pilot success criteria
- Trace recommendations back to source documents and decision steps.
- Retrieve relevant context across documents, sessions, and studies.
- Reuse prior assumptions, mappings, and protocol decisions.
- Show what changed and why it matters.
- Produce clearer, more defensible explanations for AI-assisted workflows.
Clinical AI should not start from scratch every session.
Ix gives clinical teams and AI systems persistent memory, source-backed reasoning, and decision traceability across the workflows where trust matters most.