Core SDK v2

AI-assisted authoring for governed FHIR Questionnaires

Core SDK v2 helps clinical teams create, test, review, and publish structured care plan templates without specialist development support.

Why it matters

From clinical intent to structured care plan faster

Clinically governed digital care plans are slowed by complex standards, terminology, and approval cycles. Core SDK v2 moves the first draft into the editor: the AI interprets clinical intent, proposes FHIR Questionnaire content, and keeps teams in control of review and governance.

01

Describe

Clinical teams describe the care plan requirement in natural language, without starting from code or a blank FHIR resource.

02

Generate

The Copilot prepares standards-based FHIR Questionnaire sections, items, coded choices, and reusable clinical structures.

03

Govern

Teams review, refine, validate, approve, and publish through controlled clinical governance before wider use.

Authoring capabilities

Built for clinical standards, not just form generation

The assistant is embedded in the visual editor, so generated content can be dragged, applied, validated, previewed, and governed in the same workflow.

1

Natural-language authoring for FHIR Questionnaire structures

2

Draggable generated sections and items for rapid refinement

3

SNOMED CT aligned ValueSets for coded clinical answers

4

In-place AI fixes for existing questionnaire items

5

Reusable archetype suggestions for governed clinical sections

6

Preview and validation before formal sign-off

json:questionnaire-items
[
  {
    "linkId": "vital-signs",
    "type": "group",
    "text": "Vital Signs",
    "item": [
      {
        "linkId": "blood-pressure",
        "type": "decimal",
        "text": "Systolic blood pressure",
        "required": true
      }
    ]
  }
]

Under the hood

The AI speaks the editor's native language

Core SDK v2 does not paste prose into a form builder. The Copilot streams structured responses that the editor parses into generated items or in-place fixes, preserving the visual authoring workflow while keeping the output standards-based.

Editor Copilot

A side-panel assistant streams responses into the Core SDK editor and turns generated JSON blocks into draggable items.

FHIR-aware prompt

The system prompt constrains output to supported FHIR R4 QuestionnaireItem types, Core extensions, SDC population, and extraction patterns.

Clinical terminology

Tool calls search or create ValueSets and can use SNOMED CT, helping generated choices stay coded and interoperable.

Governed reuse

Archetype search promotes reusable, approved section definitions before ad-hoc content is generated.

Accelerating clinically governed digital care plan authoring

Use AI for the first structured draft, then rely on clinical review, validation, and authorised publication before the template reaches real-world care.