Intelligence Agents¶
The clinical reasoning layers over the engines — engines are the lab; agents are the consult. The Precision Intelligence Engine is the hub that routes a case to the right agent. Each agent is decision support for a qualified clinician — never autonomous diagnosis or prescribing.
Each agent retrieves evidence, then reasons over it, then cites what it used. The citation is the part a clinician can check — an answer without a traceable source is not decision support. Retrieval quality is bounded by what has been indexed, so an agent with an unseeded corpus returns nothing rather than inventing an answer.
Agents sit above the engines, which do the deterministic computation. The single-cell pair shows the split most clearly: the engine resolves an expression matrix into cell-type clusters, and the agent explains what those clusters imply clinically.
| Capability | What it does | Domain | Status |
|---|---|---|---|
| CAR-T Intelligence Agent | RAG decision support across the five stages of CAR-T development (target identification, CAR design, vector engineering, testing, clinical), with cross-collection… | clinical | live |
| Clinical Trial Intelligence Agent | RAG decision support across 19 trial workflows spanning design and conduct: protocol design, patient matching, eligibility analysis and optimization, site… | clinical | live |
| Neurology Intelligence Agent | RAG decision support across 10 neurology domains (cerebrovascular, neurodegenerative, epilepsy, MS, neuro-oncology, movement disorders, headache, neuromuscular,… | clinical | live |
| Pharmacogenomics Intelligence Agent | RAG decision support for gene-guided prescribing across six workflows (gene, drug, profile, interaction, dosing, HLA screening): star-allele genotype to CPIC… | clinical | live |
| Precision Autoimmune Intelligence Agent | RAG decision support across 13 autoimmune conditions (rheumatoid arthritis, SLE, multiple sclerosis, type 1 diabetes, IBD, psoriasis, ankylosing spondylitis,… | clinical | live |
| Precision Biomarker Intelligence Agent | RAG decision support over multi-omics and routine labs: biological-age clocks (PhenoAge, GrimAge), 9-domain disease-trajectory risk (diabetes, cardiovascular,… | clinical | live |
| Rare Disease Intelligence Agent | RAG decision support for the diagnostic odyssey: 21 workflows (phenotype-driven search, variant interpretation, differential diagnosis, gene-therapy eligibility,… | clinical | live |
| Single-Cell Intelligence Agent | Reasoning layer over the Single-Cell Engine (Engine 8) | clinical | live |
How to read this table
Every row — including the status — is generated from the capability registry, the same file continuous integration validates. A capability marked live is served by a real model against real input, never a mock; anything planned is labelled and not shown as if it ships. See the full Capability Maturity Matrix and the Honesty Ledger.
All clinical output across these capabilities is decision support for a qualified clinician — never autonomous diagnosis or prescribing.