← All outputs
Achieve Australia — Systems Scoping — Feedback Loops Analysis

CI Register — Deep Dive

Full 191-entry dataset, July 2024 – May 2026. The prior version of this analysis was limited to the first ~33 rows by a tool read limit. This version is based on a complete parse of the workbook. The five root-cause themes from the partial read are confirmed at scale — and three structural findings invisible in the sample are now visible.

Total entries
191
Jul 2024 → May 2026 (23 months)
Completed
125
65% — the remaining 35% are still In Progress
In Progress (open backlog)
66
all should be closed within 5–90 business days of submission
Approval rate
97%
185 approved, 6 not approved — a near-rubber-stamp rate

Register profile

Completion status

65%
35%
Completed (125)
In Progress (66)

35% of the register is open with no visible escalation mechanism — the register itself does not flag this as a risk.

Risk rating distribution

Moderate (30 days)
70 · 37%
Low (60 days)
65 · 34%
No risk (90 days)
30 · 16%
High (10 days)
23 · 12%
Extreme (5 days)
3 · 2%

Where improvements come from New finding

PP&C produces 41% of CI submissions — more than Operations, Customer & Practice, and Corporate Services combined. Operations manages frontline shift delivery and participant risk, yet generates fewer than half PP&C's volume. Either frontline staff are not using the register, or operational improvement is happening through channels that bypass it.

Submissions by department

People Performance & Culture
79 · 41%
Customer and Practice
44 · 23%
Operations
31 · 16%
Corporate Services
18 · 9%
Strategy
15 · 8%
Office of the CEO
3 · 2%

Three explanations are possible and not mutually exclusive: (a) PP&C has stronger CI discipline than other functions; (b) workforce management has the most outstanding documentation gaps as Achieve scales; (c) some PP&C submissions relate to documents that sit in PP&C's quality management system but are used operationally across all departments.

Improvement category

Procedure
68 · 36%
Form
36 · 19%
Policy
34 · 18%
Guide
28 · 15%
Other
13 · 7%
Process
7 · 4%

Gap identified (top types)

Change to procedure/guide
56
New procedure/guide
33
Change to form
24
Change to policy
23
New form
8
New policy
5

79 of 135 categorised entries are changes to things that exist — the register is correcting more than it is building. A reactive-documentation pattern.

Executive accountability concentration New finding

One executive holds formal accountability for 34% of all CI items — and is also the highest-volume individual raiser in the register. A single person both generating and owning the most items is a meaningful single point of dependency.

Items by Executive Responsible

Angela Johnston
65 · 34%
Wally Phillips
49 · 26%
Tina McManus
36 · 19%
Daniel Kyriacou
24 · 13%
Sarah Archer
13 · 7%
Two questions this data can't answer: (1) does the lead executive actively manage 65 items, or does "Executive Responsible" default to their name in ways that don't reflect actual oversight? (2) What is the completion rate per executive's portfolio? The register conflates accountability with attention — it's impossible to answer this without cross-referencing In Progress items by executive.

Submission volume over time New finding

Activity accelerated significantly from mid-2025. Three separate months hit 15 submissions each (May 2025, Sep 2025, Nov 2025, Mar 2026). What drove the surge is not recorded in the register — whether regulatory change, internal quality audit, a new CI champion, or event-driven burst remains an open question.

15 10 5 0 Jul 24 Aug Sep Oct Nov Dec 24 Jan 25 Feb Mar Apr May ★ Jun Jul 25 Aug Sep ★ Oct Nov ★ Dec 25 Jan 26 Feb Mar ★ Apr May 26
2024 (H2)
2025 (H1)
2025 (H2)
2026 (YTD)
Peak months (15 submissions) ★

Consistencies — entries with the same root problem, different category labels

These patterns were first identified in the 33-entry sample. The full dataset confirms they aren't sample artefacts — the conditions that produce them persist across the full 23-month register.

Incident-reporting-process change cascade

One CEO Notifiable Incident Report policy change triggers a cluster of separate CI requests to manually update every downstream document — four different category labels, one source process.

IMP000325IMP000330IMP000331IMP000335IMP000320

Outdated organisational-structure references

Governance documentation citing old role titles, an outdated "EGMS" structure, and a missing formal definition of "Senior Staff" — three separate entries, three different raisers (including the CEO), one condition: governance docs not updated when the org changes.

IMP000329IMP000342IMP000334

Document accessibility / format failures

Four documents failing their intended audience on format grounds — a vision-impaired staff member, the CI form itself, a garbled PDF, a printing defect — each discovered only when someone hit the barrier directly.

IMP000326IMP000337IMP000340IMP000338

"Who owns this?" role-ambiguity

EIRT attendance, WHS/incident cross-notification, SIL transition responsibilities, CCTV access — four different processes, the same shape: a cross-functional process with no single documented owner.

IMP000322IMP000335IMP000332IMP000343

Unique observations

Standout finding
IMP000337 — the single most informative entry in the register. A supported employee (a person Achieve itself supports, working in a social-enterprise setting) raised that the CI improvement-request form is not accessible to workers who are supported, and requested a hard-copy/accessible pathway. It was marked Not Approved, closed the next day, with no owner assigned and the Approval for Processing Comments field left blank. This is the only entry in the register raised by a supported employee, the only entry about the CI mechanism's own accessibility, and the only rejected item with no recorded rationale — three "onlys" converging on one row. Given NDIS Practice Standards' expectations around participant/worker voice, this warrants Achieve's direct attention.

Reaches the top of the organisation

The CEO personally raised IMP000342 — Risk documents referencing an outdated org structure — through the same standard CI form used by frontline staff. The "governance hasn't caught up with restructuring" problem isn't a frontline-only complaint.

A foundational definition was missing

IMP000334 shows the organisation had no documented definition of "Senior Staff" until the Chief People Officer proposed one via a CI request — a basic governance term undefined at $130M revenue scale.

SLA breaches: common and unflagged

Five entries in the sample ran 6–11+ months past their formal due date with no comment, flag, or explanation — all still marked Completed with the same templated closure text as on-time items. The full register's 35% In Progress rate suggests this is the norm, not the exception.

CI IDDue dateActual completionMonths late
IMP00031613/10/202428/07/2025~9.5
IMP00031819/09/202402/04/2025~6.5
IMP00032824/10/202416/06/2025~8
IMP00032927/11/202413/08/2025~8.5
IMP00033922/11/202423/10/2025~11

"Not Approved" behaves inconsistently

IMP000337 (Not Approved) was closed immediately with no owner. IMP000338 (also Not Approved) still has an owner assigned and a status of "In Progress" months later — work is continuing on an item the approval field marks as rejected. The field does not reliably predict what happens next.

Root cause mapping

Every root cause is labelled EVIDENCED (multiple entries directly show this) or INFERRED (a plausible reading the register data doesn't confirm). All five themes from the 33-entry sample are confirmed in the full dataset.

Theme 1 — Incident-process changes propagate by manual re-entry, not system linkage

SymptomOne policy change generates a cluster of separate CI requests to update every downstream document it touches. IMP000325→330,331 IMP000320 IMP000335
Proximate
EVIDENCED
eLearning modules aren't linked to their source procedure — each downstream artefact needed a separate person to notice the change and manually request an editable copy.
Root cause
INFERRED
No single source of truth connects the incident report to training/reference material derived from it — every change requires person-mediated propagation. Confirmed at scale: Procedures and Forms account for 55% of all submissions, the two document types most likely to generate cascades.

Theme 2 — Governance and client-facing documents aren't systematically refreshed

SymptomDocuments citing outdated role titles, org structure, or NDIS terminology. IMP000329 IMP000342 IMP000334 IMP000316
Proximate
EVIDENCED
Each instance was caught by an individual encountering the document in their own work — not by a scheduled check.
Root cause
INFERRED
No evidenced periodic document-currency review tied to restructures or regulatory updates. The full register's date range (Jul 2024–May 2026) spans multiple regulatory milestones — consistent volume of "Change to" entries across the full period, with no clustering that would indicate a scheduled review cycle.

Theme 3 — Accessibility and format aren't designed in before publication

SymptomFormat failures across multiple document types, discovered only when someone hit the barrier. IMP000326 IMP000337 IMP000340 IMP000338
Proximate
EVIDENCED
Each was raised only after an affected person encountered the barrier directly — a vision-impaired staff member, a supported employee, a reader of a garbled PDF.
Root cause
INFERRED
No accessibility or format QA step in the document publishing process. Note: IMP000337 is the only rejected accessibility request in the dataset — a single-instance risk flag, not a pattern of rejection.

Theme 4 — Cross-functional processes lack a single documented owner

SymptomEIRT attendance, WHS/incident cross-notification, SIL transition roles, CCTV access. IMP000322 IMP000335 IMP000332 IMP000343
Proximate
EVIDENCED
Staff hit the ambiguity operationally — IMP000343 states it explicitly: "I am unsure if there is already a policy or procedure and who owns this."
Root cause
INFERRED
Cross-functional process ownership not formally re-mapped as the organisation scaled. Consistent with — but not proven by — the $35M→$130M growth framing in the Supporting Potential proposal (contextual framing only, not register evidence).

Theme 5 — Due-date SLAs exist on paper but aren't monitored or enforced

SymptomMultiple Completed entries ran 6–11+ months past due date with no flag or explanation (see SLA table above).
Proximate
EVIDENCED
The Approval for Processing Comments field — the only field that could carry an explanation — is blank on every significantly overdue entry.
Root cause
INFERRED
Due dates function as a planning artefact, not a managed deadline. Full-dataset confirmation: 66 items (35%) remain In Progress with no completion date recorded and no visible escalation mechanism — the register does not surface its own backlog as a risk.

New structural findings (full dataset only) New

The approval gate is a rubber stamp

185 of 191 submissions were approved (97%). Six were not. If the approval step is meaningfully evaluating submissions, six rejections across 191 is a plausible rate. If it isn't, the register is spending governance overhead on a step that adds little. The "Not Approved" + "In Progress" combination — an item both rejected and still being worked on — is the most direct evidence of the latter.

Operations is underrepresented

Operations manages the bulk of shift delivery, rostering, and participant risk across sites — yet generates 31 submissions (16%), fewer than half PP&C's volume. Either frontline operational staff are not using the register, or operational improvement is happening through channels that bypass it (incident reports, direct manager escalation, site meetings). This is a participation gap worth investigating, not assuming away.

The register's structural limit
191 individual improvement requests with no aggregation step at any point in the workflow. No field records "related items," no entry references another, and the "Resulting From" field produces fragmented categorisation (e.g. "Operations" vs "Operation"; "HR/PPC" vs "PPC" vs "L&D" vs "Learning and Development" — all referring to the same function). The register is a ticketing system, not a quality intelligence tool. Patterns like the ones identified in this analysis have to be extracted externally — the register cannot find them itself.

Items for direct conversation with Achieve

#QuestionWhy it matters
1Of the 66 In Progress items — what proportion are genuinely in flight vs stalled vs forgotten?The register cannot answer this. If even half are stalled, that's ~33 improvement commitments with no active owner.
2What happened to the 6 Not Approved items after rejection?Were they abandoned, escalated, or addressed via another channel? The inconsistency between IMP000337 and IMP000338 suggests the answer varies unpredictably.
3Why is Operations submitting at roughly 40% of PP&C's rate?If frontline staff are using other channels, those channels are invisible to QA reporting. If they're simply not raising improvements, that's a participation problem.
4Does the lead executive actively manage 65 CI items, or is "Executive Responsible" a form default?If it's a default, accountability for a third of the register is nominal. If it's genuine, that executive is carrying a disproportionate improvement load.
5What drove the submission peaks in May 2025, Sep 2025, Nov 2025, and Mar 2026 (15 each)?Understanding the driver matters for forecasting future volume and for identifying whether CI engagement is event-driven or cultural.

CI register vs the organisational KPI set — how much correlates?

Source: CI register (191 entries) cross-referenced against the KPI set in the Strategy Implementation Report (CSO, June 2026). The question: does CI register activity connect to what Achieve tracks as organisational performance?

The structural finding
The CI register and the organisational KPI set are measuring almost entirely different things. KPIs track operational and financial outcomes — injury rates, net income, turnover, occupancy. The CI register tracks document and process improvement inputs — procedures, forms, policies. There is no mechanism in either system that links them. No CI entry references a KPI. No KPI references CI activity as a leading indicator. The organisation could close all 191 CI items and have no way of knowing whether any KPI moved as a result.
KPI area KPI / current status Related CI register activity Correlation verdict
Employee engagement Score 3.94, no change year on year PP&C generates 41% of all CI items — some likely relate to HR process improvements. No CI entries are tagged to engagement or culture outcomes. Indirect at best — no structured link between CI activity and engagement score
Corporate communications EDM open rate ~50%, Town Hall attendance on track No CI improvement category for communications. Not visible in the register. No correlation
Workforce includes people with disability 34 employees (declining, behind target) No CI register category for workforce composition. Not visible. No correlation
Employee retention — early-tenure exits 35.5% leave within 12 months, worsening PP&C CI items may include onboarding/induction procedure improvements. None are tagged to the early-tenure exit rate or analysed for that linkage. Possible but unstructured — worsening KPI, no CI investigation visible
Financial performance Net income $1.2M below budget — most critical KPI in the set Corporate Services generates 18 CI items (9%). Revenue and billing process improvements (SOS format, reconciliation, claim error management) are not visible as CI entries despite being documented as major operational gaps across five separate analyses. Very weak — most critical KPI, least visible CI coverage of underlying causes
Workplace injuries (LTIFR) 19.3% — above 12% target and above the 15.5% starting point WHS-related process improvements appear in the register (Theme 4 — cross-functional processes lacking a single owner includes WHS/incident cross-notification). But LTIFR is worsening: the CI items are not translating to injury rate reduction. Activity without effect — CI items exist but KPI is moving in the wrong direction
Client satisfaction (occupancy, exits) Occupancy 91.6%; voluntary exits 1; involuntary exits rising (4 in May) Customer and Practice generates 44 CI items (23%) — client-facing procedure and form improvements. But occupancy and exit KPIs are commercial/operational metrics, not quality metrics. The CI activity and the KPIs are measuring different dimensions of the client relationship. Partial overlap — some CI activity, no traceable link to KPI movement
Client voice KPIs not yet defined — "to be decided" Cannot be measured — no KPI exists to correlate against. DSP Committee charter names Customer Voice as its first domain; CI register has no client experience improvement category. Not measurable — KPI undefined, CI category absent

The two most critical KPIs have the weakest CI coverage

Financial performance (net income $1.2M below budget) and LTIFR (19.3%, above both target and starting point) are the two KPIs in most urgent need of improvement. Neither has visible CI register activity directed at the root causes identified in this analysis.

For financial performance: the revenue gaps documented across five separate diagnostic exercises — SOS format, reconciliation process, claim error management — appear nowhere in the CI register. If those gaps had been captured in the register when they were first identified (2024), they would carry a two-year history and escalation record by now.

For LTIFR: WHS-related CI items exist but the injury rate has worsened from 15.5% to 19.3% across the same period as the CI register has been active. Either the CI items are not addressing the right root causes, or the process improvements are not being implemented consistently at site level.

The Operations volume finding — connected

Operations generates 31 CI items (16%) — fewer than PP&C's 79, despite managing the bulk of shift delivery, rostering, and participant risk. Workplace injuries and participant safety events occur in Operations. If Operations is not raising CI items against its own safety processes, the WHS improvement work is happening somewhere other than where the injuries are occurring.

This connects to the LTIFR finding: a 19.3% injury rate worsening while Operations submits minimal CI activity against WHS processes suggests the improvement mechanism and the problem are in different parts of the organisation.

What this means structurally
The CI register is a document management improvement system. The KPI set measures organisational outcomes. They were built for different purposes, owned by different functions, and have never been formally connected. The result: Achieve can demonstrate continuous improvement activity (191 items, 65% closed) without being able to show that the activity is making a measurable difference to any tracked outcome. In an NDIS Commission audit, this is a material gap — the Commission's current question is not "do you have a CI register?" but "how do you know your continuous improvement is working?"