AI Access — the machine door
This platform has two doors to the same curriculum. Humans use the course site. AI seats and agents use this door: a structured, machine-readable mirror of the exact same course.
The manifest
/ai/manifest.json is the single entry
point: platform identity, curriculum version, every course and module, each module's
status, and the URLs of its human page, machine-readable text, and media assets
(deck, video, job aid, narration, transcript, captions).
Per-module text
Every Beginner module's full lesson text — narration script, word tracks, floor
drill, self-checks, field assignment, and test-relevant facts — lives at
/ai/beginner/B<n>.md (e.g. /ai/beginner/B1.md),
with a metadata header (course, module, status, curriculum version).
Versioning
The manifest carries curriculum.version and an updated
timestamp. AI consumers should re-read the manifest to detect curriculum updates:
when the version or timestamp changes, re-ingest the modules whose status or assets
changed. Statuses move planned → in-production → live as production verifies each
module's media.
Management reports
/ai/reports.json is the machine-readable twin of the
management layer (/manage/): roster & progress, test results and pass
rates, the weakness board (most-missed items, miss rate by unit, per-trainee weak units and worst
stage), the retest queue with overdue/escalation flags, the certification register, and the
Goshen vs Fort Wayne rollups — all rendered from the same Academy records the human pages use.
AI consumers auditing live work can read a trainee's certified level and known weak units here
before judging his threads.
AI practice chatbots
Every Beginner module page carries AI-customer practice scenarios (Gemini)
built from that module's own lesson — a hidden buyer card, a planted trap, and a
grader pass against the module's rubric when the chat ends. Engine endpoint:
https://training.bccars.app/practice-api (health + public scenario catalog at
/practice-api/scenarios); graded sessions land in the Academy
records as practice_attempts, aggregate into
/ai/reports.json under practice, and
surface in the management practice report.
What AI consumers do with it
Ingest the curriculum, tutor salespeople against it, audit live work (threads, notes, follow-up) against its standards, and track which curriculum version any audit or tutoring session ran against.