Agentic layer

AccrediSense AI Orchestrator

One request. The orchestrator decides which specialised agents act, in what order — and pauses for human approval before anything consequential.

Specialised agents

10

Each with its own scope and autonomy

Live readiness

81%

Shared context for every agent

Approval gates

Human-in-the-loop

Consequential actions require sign-off

Voice commands

Beta

Speak to navigate and assess

Ask the orchestrator

Agent registry

Standards Agent

advisory

Interprets authorized standards and requirement scope.

Reads: Licensed standards knowledge base

Produces: Applicable requirements, evidence expectations

Evidence Agent

assistive

Analyses documents and evidence sufficiency.

Reads: Documents, evidence wallet, audits

Produces: Coverage, freshness and gap assessment

Surveyor Agent

assistive

Conducts persona-based interviews.

Reads: Requirements, prior findings

Produces: Scored interview responses

Tracer Agent

assistive

Simulates patient journeys across departments.

Reads: Pathways, departments, requirements

Produces: Step scores, observed weak points

Remediation Agent

supervised

Builds corrective action plans.

Reads: Findings, owners, history

Produces: Actions with owners and deadlines

Learning Agent

supervised

Identifies knowledge gaps and targets training.

Reads: Interviews, findings, LMS

Produces: Course recommendations and assignments

Competency Agent

assistive

Assesses staff readiness and currency.

Reads: Competency records, assessments

Produces: Reassessment lists by role

Executive Agent

advisory

Summarises organizational risk for leadership.

Reads: All agent outputs, AccrediScore

Produces: Narrative risk brief

Change Impact Agent

assistive

Analyses policy and standards changes.

Reads: Document versions, mappings

Produces: Impacted requirements, departments, staff

Research Agent

advisory

Surfaces organizational knowledge from authorized sources.

Reads: Organizational memory, past surveys

Produces: Precedents and recurring patterns

Voice-first accreditation
Beta

The voice command service is abstracted — a real speech-to-text provider plugs in without UI changes.

Agents recommend. Consequential actions — assigning large-scale training, changing records, closing findings — route through human review and are written to the audit trail.

Audio guide · purpose, benefit & ROI