How It Works

Reading the evidence

ZSF reads the Program or Project Status Report as the primary source. The diagnostic unravels the reporting — its commentary, its dimensional assessments, its footnotes, its cross-references, its attachments — and draws out the connections that shape what the evidence supports.

Where the evidence supports the reported position, the diagnostic confirms it with the source cited. Where the evidence supports a different recommendation, the diagnostic provides that recommendation with the source cited. Every claim carries its evidence.

The substrate

The diagnostic combines AI, analytical discipline, and established governance discipline. AI reads the reporting across its commentary, dimensional assessments, footnotes, and attachments. Analytical discipline applies the tolerance rules, tests dimensional aggregation, tracks recurrence across cycles, and structures the recommendations. Established governance discipline — refined over decades of practice and broad global acceptance — provides the framework structure within which all of this is consistent, honest, and defensible.

Together they form the intelligence that makes the diagnostic reliable. Three capabilities emerge from the combination.

The intelligence is joint. The recommendations are not the AI's original commentary, and not the governance discipline applied in isolation. They are what all three — AI, analytical discipline, and established governance discipline — produce together. Framework and discipline together are what makes the joint intelligence deliver.

Applying the methodology

The diagnostic applies the assurance framework adopted widely in conjunction with portfolio lifecycle frameworks. Each recommendation anchors to a specific assurance criterion drawn from this framework.

For programs and projects working under other frameworks — an organisation's own delivery framework, a sector-specific assurance discipline, or the frameworks established by professional bodies — the diagnostic expresses the same discipline in the language of the framework the work operates under. The discipline holds; the vocabulary respects the reader's context.

Government programs and non-government programs receive language calibrated to their accountability. Government programs work within assurance frameworks where compliance language applies. Non-government programs receive the same discipline framed as what sound delivery governance would apply.

Forming the recommendations

The diagnostic forms evidence-supported recommendations across five modes.

Where a dimension is reported at a position stronger than the evidence supports, the recommendation reflects what the evidence supports, with the source cited.

Where measurable thresholds are engaged — schedule variance, financial variance, risk severity — the tolerance rules apply and the recommendation reflects them.

Where established assurance discipline expects evidence the report does not provide, the absence is surfaced with the criterion it engages, so the board sees where evidence is available and where further work supports the direction.

Where the same issue appears across multiple reporting cycles, the recurrence is surfaced with the actions that address it. Repeating issues rarely resolve without direction; the board sees the pattern and the steps that break it.

Where the dimension-level assessments do not aggregate to the headline RAG under the tolerance rules, the diagnostic surfaces both — the reported headline and the position the individual dimensions support. The board reads both and directs from what the evidence supports.

Each recommendation is anchored to evidence. Each recommendation names the criterion it engages.

Matching the assessment to the work

The assessment frame matches the work being assessed. A single-project status report is assessed at project tolerance. A multi-project programme is assessed at programme tolerance. Portfolio-level reporting applies portfolio aggregation.

The frame determines what the tolerance rules apply to. Matching the frame to the work is the discipline that puts the recommendations the board needs in front of them.

The audit trail

The diagnostic records the source material read, the assessment applied, where the assessed position aligns with the reported position, where it differs, and the criteria that anchored each recommendation.

The diagnostic is itself the audit artefact. When the decision made on it is later examined, the record demonstrates that the board directed on evidence, tested independently at the point of decision.

Confidence and stability

The diagnostic records confidence per recommendation, so the board reads each recommendation knowing how firmly it is anchored. Recommendations tied to explicit source statements carry higher defensibility than recommendations derived from pattern inference. Both are honestly named.

Every recommendation carries its stability class — deterministic where tolerance rules produce the assessment directly, interpretive where the recommendation integrates pattern evidence across the report. The reader sees the reasoning behind each recommendation.