The Ribbon — Instance Anchoring & Post-Instance Intelligence
The Ribbon was introduced in Workshed + Loom = Magic as the immutable historical record — frozen instances that can never be rewritten by future template or rate changes. The Core Engine Schema implements this through the lock endpoint and denormalised instance_* tables.
But the Ribbon is more than a ledger. A locked instance is a universal anchor point — a UUID that every operational artifact attaches to. And once artifacts are anchored, the system can do something no manual process can: automatically validate, cross-reference, and learn from what actually happened.
This document describes what hangs off a locked instance, what the system can do with those attachments, and how automated post-instance intelligence turns operational data into organisational insight.
The Instance as Universal Anchor
Every locked instance carries a UUID. That UUID is the foreign key for everything that happened during that instance:
| Attached To | What It Contains |
|---|---|
instance_attendance | Who attended, arrival/departure times, absence reasons |
instance_shifts | Which staff worked, shift start/end, break times |
instance_staff_checkin | Day-of staff check-in confirmations |
instance_vehicle_dispatch | Which vehicles were used, odometer readings, fuel |
instance_billables | Billing lines with denormalised rates, hours, totals |
instance_routes | Transport route cards with pickup/dropoff sequences |
instance_timeslots | The actual timeline cards as they ran on the day |
instance_artifacts | Shift notes, incident reports, photos, documents |
instance_tags | Semantic tags with vector embeddings for search |
Standard tags and UUIDs mean any system — the admin portal, Reggie, the billing engine, the compliance layer — can reach into a locked instance and find everything relevant to that moment in time.
This is not a filing system. It is a structured memory of what happened, anchored to a specific program on a specific date, with every participant, staff member, vehicle, billing line, and narrative artifact attached.
Post-Instance Feedback Loops
Once an instance is locked and its artifacts are attached, the system can initiate automated feedback collection. None of this requires human scheduling or follow-up — the lock event itself is the trigger.
Participant Feedback
After a shift completes, the system can SMS participants (or their carers/guardians) asking:
- Did you enjoy the activity today?
- Was everything as expected?
- Any changes or problems you'd like to report?
- Would you rate today's experience? (1-5)
Responses are parsed by Reggie, tagged with sentiment, and attached to the instance as artifacts. Over time, this builds a participant-level satisfaction timeline that is anchored to specific programs, staff, and venues — not vague quarterly surveys.
Staff Feedback
The system can SMS staff after their shift:
- How did the shift go?
- Any incidents or concerns?
- Anything you need support with?
Staff responses are equally anchored to the instance. This creates a dual-perspective record — what the participant experienced and what the staff member experienced, for the same event on the same day.
Why Automated Collection Matters
Manual feedback collection is sporadic, biased toward problems, and disconnected from the operational record. Automated post-instance collection is:
- Consistent — every instance triggers it, not just the ones someone remembers to follow up on
- Anchored — responses are attached to the exact instance, not floating in an inbox
- Timely — collected within hours, while memory is fresh
- Comparable — same questions, same structure, across all instances, making trend analysis possible
Automated Billing Validation
Locked instances contain denormalised billing lines — the NDIS code, the unit price, the hours, the ratio, the total, all as they were on the day. This creates a surface for automated validation that would be impossible with live foreign keys.
Code Validation
Each billing line's NDIS code can be checked against:
- The current NDIS price guide (is this code still valid?)
- The participant's service agreement (is this code authorised for this participant?)
- The program's approved billing codes (does this code belong on this program?)
Historical Comparison
Because every past instance for the same program has the same structure, the system can compare:
- Are the billing lines for this instance consistent with previous instances of the same program?
- Has the total changed significantly? If so, why — different attendance, different rates, or an error?
- Do the hours billed match the actual shift duration from
instance_shifts?
Payment Request Verification
When payment requests are generated from locked billables:
- Do the requested amounts match the locked line totals?
- Do the participant details match the locked attendance record?
- Does the service date match the instance date?
- Have confirmations been received for previous payment requests on this program?
Discrepancies are flagged automatically. The billing team reviews exceptions rather than checking every line.
Cross-Referencing Artifacts for Accepted Reality
A single instance may have multiple authored artifacts — a shift note from the team leader, a shift report from a support worker, an incident report, participant feedback, staff feedback. These are different perspectives on the same event.
Outlier Detection
When multiple staff submit shift reports for the same instance, the system can compare:
- Do the narratives align on key facts (activities completed, incidents, participant behaviour)?
- Are there contradictions (one report says the outing went well, another describes a major incident)?
- Is one report significantly more negative or positive than the others?
Outliers are not automatically wrong — but they are worth reviewing. The system flags them for a coordinator to assess whether the discrepancy reflects a genuine difference in experience, a misunderstanding, or an accuracy issue.
Triangulation with Feedback
Staff-authored artifacts can be cross-referenced with participant/carer feedback:
- A shift note says "great day, everyone enjoyed the park" — does the participant feedback agree?
- A shift report describes a difficult behaviour incident — did the participant's carer mention anything?
- A staff member reports everything was fine — but the participant rated the day 1/5
This triangulation builds a more complete picture than any single source. It moves the organisation from accepting the first narrative written to constructing an accepted reality from multiple perspectives.
Sentiment Analysis and Staff Wellbeing
This is where the Ribbon becomes genuinely protective of the people who do the work.
The Problem with Emotional Labour
Disability support, aged care, and community services involve emotionally intense work. Staff regularly navigate challenging behaviours, medical emergencies, family grief, and participant distress. Over weeks and months, this accumulates.
The traditional signal that a staff member is struggling is when they resign, go on extended leave, or have a performance issue. By then, it's too late for early intervention.
Detecting Drift Through Artifacts
Staff produce artifacts constantly — shift notes, shift reports, incident reports, feedback responses. These are authored in their own words, in their own emotional state, at the time of the event.
Sentiment analysis across a staff member's artifacts over time can detect drift:
- Language becoming more negative, more terse, or more detached
- Increasing frequency of words associated with frustration, exhaustion, or hopelessness
- Shift notes that were once detailed and engaged becoming minimal and perfunctory
- A pattern of negative incident framing that diverges from peer reports of the same events
What Drift Means (and What It Doesn't)
Drift does not mean the staff member is wrong. It does not mean their reports are inaccurate. It does not mean they are underperforming.
Drift means this person may be carrying more emotional weight than they can sustain. The appropriate response is support, not scrutiny:
- A check-in conversation with their supervisor
- An offer of Employee Assistance Program (EAP) access
- A temporary change in roster to reduce exposure to high-intensity participants
- Recognition that the work they are doing is hard and they are not invisible
The system detects the signal. A human decides the response.
Why This Matters for Narrative Accuracy
When a staff member is burning out, their authored artifacts may unconsciously skew negative. A participant who was "a bit unsettled" becomes "aggressive and unmanageable". A day that was "mixed" becomes "terrible".
This is not dishonesty — it is the natural result of emotional exhaustion colouring perception. By detecting drift, the system can:
- Flag that this staff member's recent artifacts may reflect their emotional state more than the objective reality
- Cross-reference against other perspectives (peer reports, participant feedback) to construct a balanced view
- Ensure that a participant's record is not permanently coloured by one burned-out staff member's worst week
- Ensure the staff member gets the support they need before the situation escalates
The Ribbon makes this possible because every artifact is anchored to a specific instance, authored by a specific person, on a specific date. Trend analysis requires that structure.
Compliance and Audit Automation
Locked instances with complete artifact attachment enable automated compliance checking:
Completeness Checks
For each locked instance, does it have:
- Attendance record with actual arrival/departure times
- At least one shift note or shift report
- Staff check-in confirmations
- Vehicle pre-trip inspection (if transport was used)
- Incident reports (if incidents were recorded)
- Billing lines matching the attendance
Missing artifacts are flagged. A coordinator can follow up while the event is recent rather than discovering gaps during an audit months later.
NDIS Audit Readiness
NDIS audits require evidence that:
- Services were delivered as claimed
- Billing matches actual service delivery
- Participants were consulted about their experience
- Incidents were reported and managed appropriately
- Staff had appropriate qualifications and checks
Every one of these requirements maps to data already anchored to the locked instance. The audit is not a scramble to reconstruct the past — it is a query against structured, anchored, immutable records.
Random Sampling
The Ribbon supports automated random sampling:
- Select N locked instances from the past quarter
- For each, verify artifact completeness, billing accuracy, and feedback alignment
- Generate a compliance score and flag any instances that need review
This can run continuously rather than annually, turning compliance from a periodic event into an ongoing quality signal.
The Intelligence Feedback Loop
The Ribbon feeds forward into the Loom. What the system learns from locked instances improves future projections:
| Ribbon Signal | Loom Response |
|---|---|
| Participant consistently absent on Mondays | Flag Monday instances, suggest intent to remove |
| Billing lines frequently adjusted after lock | Review rate setup on the template |
| Staff member's sentiment declining | Suggest roster adjustment before burnout |
| Vehicle breakdowns correlating with mileage thresholds | Flag vehicles approaching maintenance intervals |
| Participant feedback low for specific program/venue combinations | Surface to coordinator for program review |
| Shift duration consistently exceeds template slot times | Suggest template time adjustment |
The Ribbon is not just history. It is the training data for organisational intelligence.
Summary
The locked instance is the atomic unit of operational truth. Everything attaches to it. Everything can be queried from it. And from the pattern of thousands of locked instances over time, the system builds a picture of what the organisation actually does, how well it does it, and where the people doing the work need support.
| Layer | What It Does |
|---|---|
| Anchor | Every artifact, billing line, attendance record, and feedback response attaches to a locked instance UUID |
| Validate | Billing lines checked against NDIS codes, service agreements, and historical patterns |
| Cross-reference | Multiple perspectives on the same event compared for consistency |
| Detect | Sentiment drift in staff artifacts flagged as a wellbeing signal, not a performance issue |
| Comply | Completeness checks, random sampling, and audit-ready structured records |
| Learn | Ribbon patterns feed forward into Loom projections and operational recommendations |
The Ribbon is where past truth lives. But it is not a dead archive. It is an active intelligence surface that protects participants, supports staff, validates billing, and ensures the organisation can always explain what happened and why.