AI Field Manual
FM-PM-02 · Guide

Project Management  ·  Fundamentals

Assemble the weekly status report from scattered updates

It's Thursday afternoon. Seven workstream leads have sent updates in four formats, one hasn't replied, and you need three versions by tomorrow, one for the steering committee, one for the team and one for the client. Most of the work is collation. The part that matters is the judgment about whether the project is actually on track, and that part stays with you.

Status
Published
Reading time
About 13 minutes
You'll need
This week's team updates, last week's report, the milestone plan or schedule, the risk and issue log, and an approved AI assistant
Last reviewed
November 2026
Before you start

What AI is for here, and what it isn't

A status report does two jobs: it consolidates facts, and it makes a judgment about those facts. AI is genuinely good at the first job. It can merge seven updates into one structure, spot what changed since last week, notice that one workstream's date contradicts another's, and rewrite the same facts for three audiences.

It should never do the second job. The status color, whether the project is green, amber or red, is a professional judgment that depends on things the model can't see: how much float is really in the schedule, whether a vendor's "on track" means anything, what the sponsor will tolerate. A model asked to "assess the status" will produce a confident color based on tone. A project that's quietly slipping usually sounds fine in its updates. That's the whole problem.

Step 1

Fix the report's structure and status definitions

Before collating anything, write down two things that should change rarely.

The structure. Use the same sections every week so readers know where to look and changes stand out: overall status, what changed, milestones, risks and issues, decisions needed, and next week.

Your status definitions. Write down what each color means on this project, in measurable terms. For example:

Status Meaning on this project
Green All milestones in the next 30 days are forecast on time; no open issue threatens a committed date or budget line.
Amber A milestone is forecast late but can be recovered inside its float, or a risk needs a decision within two weeks to protect a date.
Red A committed milestone or budget line will be missed without a sponsor-level decision or extra resources.

Agree these with the sponsor once. They turn a gut-feel color into something you can defend, and they're what you'll compare the AI's facts against.

Step 2

Collect updates in a consistent shape

Most of the pain in status reporting comes from updates arriving in different shapes. Ask every lead for the same four things:

  1. Done this week, against what was planned.
  2. Planned next week.
  3. Dates at risk, with the milestone and a forecast date.
  4. Blockers and decisions needed, with who can unblock them.

You won't always get them. When an update is missing or vague, don't let the model fill it in. Mark it "no update received" or "unclear," and chase the lead. A status report that hides a missing update is less honest than one that shows it.

Step 3

Consolidate with AI, and find the contradictions

The status consolidation prompt
You are helping a project manager consolidate weekly updates into a
status report draft. I decide the overall status. Do not assign or
suggest a status color for the project or any workstream.

PROJECT: [name]          REPORTING PERIOD: [dates]
STATUS DEFINITIONS: [paste your Green/Amber/Red definitions]
LAST WEEK'S REPORT: [paste]
MILESTONE PLAN: [milestone, baseline date, current forecast]
THIS WEEK'S UPDATES: [paste each lead's update, labelled by workstream]

PRODUCE:
1. WHAT CHANGED since last week: completed items, new risks and
   issues, date changes (old date → new date, with the reason given).
2. MILESTONES: each with baseline, last week's forecast, this week's
   forecast, and the update that supports the forecast.
3. CONTRADICTIONS: places where updates disagree with each other, with
   last week's report, or with the milestone plan.
4. GAPS: missing updates, vague statements ("on track", "progressing")
   with no evidence behind them, and dates with no forecast.
5. DECISIONS NEEDED: what, from whom, by when, and the consequence
   of no decision.
6. FACTS RELEVANT TO STATUS: list the facts that bear on each status
   definition. Do not draw the conclusion.

RULES: Quote or cite the source update for every claim. Do not soften
bad news or add reassurance. Do not invent dates, percentages or causes.

The contradictions and gaps sections are where the value is. "Data migration: on track" next to "Testing can't start until migration delivers the test dataset, now expected 2 weeks late" is the kind of thing a busy reader of seven separate emails misses.

Step 4

Set the status yourself

Now compare the facts against your definitions and decide. Write one or two sentences that justify the color in terms of the definitions. That sentence is the most important line in the report.

  • Weak: "Amber: some delays in data migration."
  • Strong: "Amber: the test dataset is forecast two weeks late, which consumes all remaining float before user acceptance testing. Go-live (3 March) holds if migration delivers by 27 January; a decision on adding a second migration engineer is needed by 20 January."

Watch for watermelon reporting: green on the outside, red on the inside. If every workstream is green but the milestone forecast is slipping, the overall status isn't green. And don't let the color drift to protect anyone. Raising amber early, with a recovery plan, is how you avoid having to announce red later.

Step 5

Produce the audience versions

Once you've written the consolidated report and set the status, have the AI adapt it. The facts must stay identical; only the depth and framing change.

The audience adaptation prompt
Below is the confirmed status report. Produce three versions. Keep every
fact, date and status color identical across versions. Do not add,
remove or soften any risk; only change the level of detail.

1. EXECUTIVE (5 lines max): status and why, the one or two decisions
   needed with deadlines, and the date that matters most.
2. TEAM: full detail by workstream, next week's priorities, and what
   each lead needs from others.
3. CLIENT: [paste the rules for what may be shared with the client].
   Exclude internal staffing, vendor commercial terms and anything
   marked internal. Flag any fact you were unsure whether to include.

Then list, for each version, every fact you omitted and why, so I can
check nothing important was dropped.

That last instruction protects you from the most common adaptation failure: an important risk quietly disappearing from the shorter version.

Step 6

Worked example: the green report that wasn't

Fictional project. A finance system upgrade, week 14. The updates say:

  • Data migration: "Good progress, on track overall. Some issues with legacy vendor records, working through them."
  • Integration: "Interfaces built. Waiting on migrated vendor data to start end-to-end testing."
  • Training: "Materials 80% done, on schedule."
  • Last week's report: end-to-end testing to start 13 January.

A model asked to "summarize the status" writes: "Overall green. Strong progress across workstreams; minor data issues being resolved."

The consolidation prompt, which forbids a color and asks for contradictions, instead reports:

Contradiction: Integration can't start end-to-end testing without migrated vendor data. Migration describes unresolved issues with exactly those records but gives no forecast date. Last week's report has testing starting 13 January.

Gap: Migration's "on track" has no supporting forecast. Training's "80%" has no baseline (80% of what was planned by now, or of the total?).

Decision possibly needed: none stated, but the testing start date depends on the migration forecast.

The project manager calls the migration lead, learns the vendor records need a two-week cleanup, and reports amber with a specific recovery path and a decision deadline. Nobody is surprised in February.

Step 7

Send it, and keep the history

Send on the same day each week. Keep every report, because the history of forecast dates is often the clearest evidence of how a project is really going: a milestone that's moved six times in eight weeks tells a story no single report does.

Weekly status report template
[Project] — Status report, week ending [date]
Overall status: [GREEN / AMBER / RED]
Why: [one or two sentences, in terms of the status definitions]

DECISIONS NEEDED
- [decision] — from [name] — by [date] — if not decided: [consequence]

WHAT CHANGED SINCE LAST WEEK
- Completed: [items]
- Dates moved: [milestone] [old date] → [new date] — reason: [reason]
- New risks / issues: [items]

MILESTONES
Milestone          Baseline    Last week   This week   Status
[name]             [date]      [date]      [date]      [on track / at risk / late]

TOP RISKS AND ISSUES
- [risk/issue] — owner [name] — response [action] — due [date]

NEXT WEEK
- [priorities]

Updates not received: [workstreams, or "none"]
AI assistance: consolidated with [tool]; status set by [your name].

What stays with you: the status color. AI doesn't get to call a project green.

Esc

Saved in this browser

Reading list

Contents

Reference

Keyboard shortcuts

/ or Ctrl K
Search the manual
↑ ↓
Move through search results and index entries
Enter
Open the selected result
← →
Switch discipline tabs; previous / next entry in a workflow card
Esc
Close any dialog
?
Show this list