Practical guide

How to create a programme status report with AI

A useful programme status report helps a named governance body make decisions. It does not reward a programme for producing a polished summary or hide weak evidence behind a traffic-light colour. AI can help assemble controlled inputs, compare versions, find omissions, and draft concise explanations. People remain accountable for status judgments, forecasts, escalation, and approval. This guide.

By Two Prune

A useful programme status report helps a named governance body make decisions. It does not reward a programme for producing a polished summary or hide weak evidence behind a traffic-light colour. AI can help assemble controlled inputs, compare versions, find omissions, and draft concise explanations. People remain accountable for status judgments, forecasts, escalation, and approval. This guide uses the TRACE method: terms, results, assumptions, changes, and escalations.

Source basis and limits

The current UK Government Functional Standard for Project Delivery sets expectations for programme governance, planning, control, roles, solution delivery, transition, use, and disposal. The Orange Book describes risk management as an integral part of governance and decision-making and emphasizes structured communication of threats, opportunities, and uncertainty. The AQuA Book calls for traceable records of data, assumptions, decisions, limitations, changes, and assurance. These sources support a decision-oriented status report; they do not prescribe one universal template or validate a particular programme's status. GovS 002: Project Delivery The Orange Book The AQuA Book

Sources: [1] [2] [3]

T — Terms and reporting contract

Start with a reporting contract approved by the programme's governance body. Record:

Define colours or labels before seeing the results. A red status should mean a specific tolerance, control, dependency, or outcome condition has been breached or is forecast to be breached. It should not mean the writer feels worried. Green should not mean “no one reported a problem.” Add a confidence label so weak evidence cannot appear stronger merely because the status is green.

Use not available, not applicable, and not assessed distinctly. Do not convert missing evidence into zero risk or unchanged status.

  • the decision audience and meeting date;
  • reporting period and data cut-off;
  • scope, outcomes, benefits, milestones, dependencies, and tolerances;
  • definitions for status, trend, confidence, exception, and escalation;
  • source owners and refresh dates;
  • author, independent reviewer, programme owner, and approver; and
  • what belongs in appendices rather than the decision summary.

R — Results and outcome evidence

Report progress through decision-relevant evidence, not activity volume. For every outcome or benefit, show the baseline, current observation, target or tolerance where approved, period, population, source, owner, and interpretation limit.

| Outcome or result | Baseline | Current evidence | Target or tolerance | Trend | Confidence | Owner | Decision implication | | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | | | | |

Separate at least four layers:

A milestone completed on time is not evidence that a benefit occurred. A training attendance count is not evidence of applied capability. If an outcome needs more time or a different evaluation design, state that boundary.

  • delivery: completed work and accepted outputs;
  • operational readiness: people, process, technology, controls, support, and transition;
  • use and quality: whether intended users can use the solution and whether outputs meet stated criteria; and
  • outcome or benefit: the downstream change the programme seeks.

A — Assumptions, risks, issues, and decisions

Maintain a compact decision-risk register tied to the report.

| ID | Type | Statement | Evidence | Exposure or consequence | Owner | Response | Trigger or due date | Status confidence | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | | | | | |

Distinguish:

Give each item a testable statement, source, owner, response, and trigger. Avoid copying an unchanged top-ten list for months. Explain what new evidence changed likelihood, consequence, timing, or confidence. Preserve material minority or specialist concerns even when the overall status is favorable.

  • an assumption that the plan currently relies on;
  • a risk that may occur;
  • an issue that has occurred;
  • a dependency controlled outside the programme;
  • a decision awaiting authorized judgment; and
  • an opportunity that could improve value or resilience.

C — Changes, forecast, and control evidence

Compare the approved baseline with the current forecast. Show changes to scope, schedule, cost, benefits, quality, resources, dependencies, and risk exposure. For every material change, record the previous value, current value, reason, evidence, decision authority, and downstream effect.

Forecasts should show a range or scenario when one date or cost would imply false precision. Name the assumptions driving the forecast and the point at which the programme would need to replan, pause, reduce scope, or escalate.

AI can compare registers, minutes, plans, and prior reports to propose changed items. A human must reopen the authoritative record and verify every material claim. Never ask AI to reconcile conflicting figures by choosing the most plausible value. Preserve the conflict and route it to the named owner.

E — Escalation and next decisions

End with the decisions required from the governance body, not a generic request for awareness.

| Decision needed | Options | Recommended disposition | Evidence and uncertainty | Decision owner | Deadline | Consequence of delay | | --- | --- | --- | --- | --- | --- | --- | | | | | | | | |

State whether the recommendation is to continue, repair, narrow, pause, stop, rebaseline, investigate, or accept a bounded risk. Link each recommendation to the evidence in the report and show what would cause it to be revisited.

The programme owner should approve the report's status and escalations. Functional specialists approve only within their mandates. The report should not imply that a programme manager can accept legal, privacy, security, financial, accessibility, or operational risk on behalf of the accountable owner.

Bounded AI workflow

Use AI in controlled passes:

Retain a report version, source snapshot, calculation record, decision log, and assurance note under the organization's information-handling and retention rules.

  • Inventory: list approved source files, owners, dates, periods, and missing items.
  • Reconcile: compare stable IDs and flag conflicting values without resolving them.
  • Draft: write one TRACE section at a time with source IDs beside material statements.
  • Challenge: search for stale evidence, hidden assumptions, optimistic language, unsupported causality, and status without a rule.
  • Verify: reopen source records, reproduce calculations, obtain specialist review, and record accepted and rejected changes.

Worked report: a fictional service-transition programme

A programme is replacing a manual case-allocation process. The prior report is green because software development is 92% complete and training invitations were sent. TRACE changes the decision picture.

The reporting contract defines readiness as accepted interface tests, accessible end-to-end task completion, trained support coverage, a rehearsed fallback, and signed data controls. Evidence shows the build milestone is on plan, but two source systems have not completed reconciliation, keyboard-only testing found a blocking defect, and weekend support coverage has no owner. The benefit measure—reduced misrouting—cannot yet be observed in live service.

The programme records one issue, two dependencies, and an assumption that the legacy queue can remain available for six weeks. The AI comparison draft incorrectly says reconciliation is “nearly complete” because one source uses a 96% record count. Human review finds that the unmatched records are concentrated in a high-consequence case type. The report removes the aggregate reassurance and preserves the exception.

The overall delivery status remains on plan, operational readiness is red under the pre-agreed rule, and confidence is medium because one supplier's evidence is dated. The governance request is not “note progress.” It asks the senior responsible owner to choose between a two-week transition delay and a narrowed launch excluding the affected pathway. Each option shows cost, service consequence, fallback, uncertainty, and the evidence needed before reconsideration.

Report verification checklist

[ ] Audience, period, cut-off, scope, definitions, tolerances, and owners are explicit.

  • [ ] Delivery, readiness, use and quality, and outcomes are reported separately.
  • [ ] Every material figure has a source, owner, period, population, and calculation route.
  • [ ] Missingness and evidence confidence remain visible.
  • [ ] Assumptions, risks, issues, dependencies, opportunities, and decisions are distinguished.
  • [ ] Baseline changes have authority and downstream effects recorded.
  • [ ] Critical exceptions are not averaged away.
  • [ ] AI-generated statements, comparisons, calculations, and citations were independently checked.
  • [ ] The final section asks for named decisions with options, deadlines, and consequences.
  • [ ] Limitations, unresolved conflicts, and revisit triggers are preserved.

Two Prune boundary

Two Prune can assess whether a participant interprets a programme brief, navigates supplied evidence, uses AI thoughtfully, reconciles records, detects unsupported claims, controls uncertainty, and writes a decision-ready status report in a realistic simulation. It reports observed assessment performance. It does not run programme governance, certify delivery status, accept risk, predict programme outcomes, or make employment decisions.

Sources

UK Government Functional Standard GovS 002: Project Delivery

  • HM Treasury: The Orange Book
  • GOV.UK: The AQuA Book

Sources: [1] [2] [3]

Sources

  1. 1.GovS 002: Project Delivery · UK Government
  2. 2.The Orange Book · UK Government
  3. 3.The AQuA Book · UK Government

Explore this topic