SLAtech Medical
88/100BAA-eligible, FHIR-conformant, polished Hebrew RTL
Medical eval scoreboard
SLAtech Medical: 88/100
Reproducible 200-question Med-specific eval harness. +17-point lift vs generic SLAtech-Business (71/100). Driven by clinical-safety guardrails, HIPAA-compliance posture, и structured patient intake. Pairs с umbrella eval scoreboard, Med glossary и Med FAQ.
Score breakdown by category
| Category | Med-tuned | Generic | Lift |
|---|---|---|---|
| Clinical-safety guardrails Symptom-triage queries routed к human-handoff где clinical advice would be UPL-adjacent. Generic chatbots attempt direct diagnosis (failure). |
92 | 64 | +28 |
| Patient intake quality Structured intake captures reason для visit, insurance, allergies, medications. Generic chatbots dump intake к unstructured free-text. |
89 | 73 | +16 |
| HIPAA compliance посture PHI redaction at ingest, BAA-eligible single-tenant option, audit-log per-action. Generic chatbots не ship PHI redaction. |
91 | 58 | +33 |
| FHIR / EMR integration queries FHIR Patient / Appointment / Practitioner / Encounter resources. Generic chatbots can't quote EMR slot availability. |
86 | 67 | +19 |
| Multilingual clinical (HE / RU) Generic chatbots actually score higher here due к broader auto-translate coverage. Med-specific terminology в Hebrew / Russian is а continuing investment area. |
84 | 92 | -8 |
Competitor comparison
Intercom Fin (generic)
73/100Not BAA-eligible by default, English-first, no FHIR integration
Ada (mid-market enterprise)
78/100SOC 2 Type II но weaker FHIR integration, implementation-consultant required (6-12 weeks)
Tidio Lyro (generic SMB)
65/100No HIPAA compliance, no Hebrew RTL polish, conversation cap on lower tiers
Reproduce the eval against your own tenant
Eval methodology is open-source. 200 sealed Med-specific questions с LLM-as-Judge scoring on factuality, hallucination и confidence axes.