The Six-Agent Wound Visit: What Runs in Parallel to Draft One Chart Credit

One credit doesn't buy one model call. It buys a coordinated run of six AI agents working the same wound visit in parallel. Here's what each one does and how they hand off.

Agent 1 — The Scribe

Listens to the encounter audio and produces a structured narrative: history of present illness, patient-reported symptoms, and provider observations. Distinct from a generic ambient scribe because it's tuned to wound care vocabulary — undermining, tunneling, exudate character, periwound condition. See the AI scribe for wound care for the standalone workflow.

Agent 2 — The Imaging Agent

Processes point-and-capture photos into calibrated measurements: length, width, depth, area, and tissue composition percentages. No manual rulers, no separate app. Powered by AI-powered wound imaging.

Agent 3 — The Chart Builder

Merges the scribe narrative and imaging output into a compliant SOAP note. Populates assessment, plan, and wound-specific fields required for downstream billing and audit defense.

Agent 4 — The Coder

Suggests ICD-10, CPT, and HCPCS codes aligned to the documented work. Flags documentation gaps that would cause denials before the chart is signed — not after.

Agent 5 — The Healing Analyst

Compares this visit's measurements to prior visits, computes healing trajectory, and surfaces stalled wounds via the healing dashboard. This is the agent that turns documentation into a longitudinal clinical signal.

Agent 6 — The Educator

Generates patient-specific education materials tied to the wound type, care plan, and comorbidities — offloaded from the clinician's post-visit workload.

Why parallel matters

Run sequentially, six agents would take 20+ minutes. Run in parallel, the chart is ready for review in about five. The clinician approves — they don't retype. That's what one credit pays for.