Transformation oversight you can defend.
Figures from an uploaded file checked against the pack, traced to that file, and every accepted exception justified in writing — before the SteerCo. Generated from your Monday or Jira export — one upload, not a rebuild.
Data Migration: Green in the SteerCo, Red in its own WG Report.
R2 · affects Exec Summary, slide 2
The same workstream cannot carry two RAGs. One of them is the truth — and the pack doesn’t get to pick. Flagged before the pack renders.
- Deterministic rule engine — not an LLM's opinion
- Live rule coverage published in-app
- Built for FCA-regulated governance
- Append-only · audit-grade
Sound familiar?
Figures you can't defend
A board member asks where a number came from, and the honest answer is "the last deck."
AI tools that make things up
Generic copilots invent a confident sentence with nothing underneath it.
Hours rebuilding packs
Every cycle, the same figures retyped by hand across six artefacts that then disagree.
For the Transformation Director / EPMO Lead
You answer to the board, the CFO and the regulator. PMOpack means the figures from your uploaded file have been checked against the pack and every accepted exception justified before you present — no surprises, no 11pm reconciliation, no “let me come back to you on that.”
For the Head of PMO / Programme Manager
Stop re-checking your team’s numbers by hand. PMOpack verifies the figures across every artefact, so your people spend time on delivery, not cross-referencing spreadsheets.
The templates are already written — by 17 years of delivery.
Stand up your own PMO tooling and you're maintaining a system by hand on top of your actual job. PMOpack gives you a board-ready SteerCo pack from data you already have. No formula to break. No system to build.
It checks the pack, not the page.
PMOpack checks each artefact against the rest of the period’s pack before it renders — the same way, every time.
Every registered check is listed inside the product, on a coverage page rendered live from the rule registry the engine runs — so it can never disagree with the engine.See the live rule coverage (sign-in required)
If your SteerCo says 1 September and your Milestone Report says 1 October, that’s a CRITICAL flag.
No reviewer catches that across six documents at 11pm. The engine does, every time — before either page reaches the board.
“…residual exposure of£50Kpending mitigation.”
If your risk slide mentions £50K, we check the £50K is real.
PMOpack reads the money figures quoted in your narratives and reconciles each one against the Budget Tracker. A number quoted in prose with no approved budget home gets flagged — the phantom figure buried in a sentence.
Every figure from an uploaded file has a paper trail — and, where the PM has marked it verified, who signed it off and when.
Trace any number back to the file you uploaded and the PM’s verification note, where one has been recorded — see exactly which document backs it and, where the PM has marked it verified, who signed off and when. Not “we used your file” — “this file is verified, by this person, on this date.”
One red critical-path task turns the workstream red. That’s not a setting.
Generic tools roll up status by volume. PMOpack rolls up by delivery reality — a single red task on the critical path drives the whole workstream red, before any percentage threshold. Seventeen years of delivery says that’s how it actually works.
Three steps. No rebuild.
- 1
Upload verified sources
Excel, CSV, a Monday.com export, or a Jira export. Flag each as verified with a note.
- 2
The engine checks it
Deterministic coherence checks, gap detection, a completeness score and a path to green.
- 3
Artefact with its record
Figures from an uploaded file traced to source, integrity-badged, ready to download.
You confirm every mapping first
Every column mapping is shown to you for review — nothing is written to an artefact until you confirm it. Once you do, your confirmed mappings and the transformations applied are saved with the snapshot.
Every figure from an uploaded file keeps its source file and field, and every transformation applied to it — including the row, where the parser captured one.
Every column mapping is shown to you for review — nothing is written until you confirm it.
Once you confirm, the journey — source, transformations applied, confirmed mapping — is saved with the snapshot. Cancel the review and nothing is written.
Four questions the board will ask. Answered before they ask them.
Not a spell-check on your prose. A reconciliation across the whole reporting pack — run the same way every time, on the figures from an uploaded file, before anything renders.
Every date, reconciled across every artefact
Go-live, milestone sets and period ends checked against each other — plus weekends, public holidays and your declared blackout periods.
Figures from an uploaded file traced back to the budget
Workstream totals, forecast at completion and contingency draws reconciled — including figures quoted in prose, where no reviewer looks.
A RAG that matches the detail underneath it
Programme and workstream status reconciled against the register beneath it, with a path to green required on every amber and red.
Decisions that actually happened
Approved decisions carried into this period's action log, each with a named owner and a due date. One recorded as not enacted is raised, not buried.
Eight artefacts. Each badged honestly.
Everything marked Live generates now — from validated, source-attributed, PM-attested data, integrity-checked before it leaves the room. Anything marked Coming would stay honestly labelled until the day it generates.
- SteerCo PackLiveIntegrity-checked ✓SteerCo PackDecision-required board pack
- Status ReportLiveIntegrity-checked ✓Status ReportRAG reconciled to source
- RAID LogLiveIntegrity-checked ✓RAID LogPath-to-Green enforced
- Milestone ReportLiveIntegrity-checked ✓Milestone ReportBaseline vs forecast
- WG ReportLiveIntegrity-checked ✓WG ReportWorking-group status, escalated
- Risk Heat MapLiveIntegrity-checked ✓Risk Heat MapReconciled to the RAID register
- Plan on a PageLiveIntegrity-checked ✓Plan on a PageOne-page programme roadmap
- Budget TrackerLiveIntegrity-checked ✓Budget TrackerForecast vs approved, reconciled
A useful deck is not yet a defensible reporting pack.
Use AI creation tools where they help. PMOpack starts where the advisor must stand behind the result for a client, a board and the next reporting cycle.
Why not just use…?
Some of these are excellent products, and a few solve a problem we do not solve at all. Here is the comparison written so their own teams could read it.
Why not just use Shibumi?
Shibumi is a serious enterprise platform — zero-code strategic portfolio management, in the Gartner Magic Quadrant for the sixth consecutive year, holding initiatives, risks, financials and benefits in one live model. Because a dashboard and a board presentation render the same attribute from that one model, a figure is not re-keyed twice. PMOpack works the other end of the problem: on the separate governance documents a programme actually publishes, reconciling the same figure across them and holding the pack back while a critical inconsistency is unresolved.
Why not just use McKinsey Wave?
Wave is real software, not a consulting deliverable — sold on AWS Marketplace, self-manageable by McKinsey's own account, and used by some clients for over a decade. It names aggregation and reconciliation of metrics as a capability, runs finance-verified stage gates and keeps audit trails. Like Shibumi, it prevents divergence by holding one datastore. PMOpack detects it across documents authored separately, and blocks publication until each critical flag is fixed or accepted with a written justification the database will not let you skip.
Why not just use Atlassian Focus?
Focus has moved fast. Since Funds went generally available in June 2026 it carries budgets, forecasts, actuals, baselines and variance beside the goals and work, and Rovo will draft a structured focus-area update. It is a live command centre, and Atlassian is clear it is not a financial system of record. PMOpack produces the published artefact itself, and checks the numbers across every artefact in the period before it lets the pack out.
Why not just use Copilot, Rovo or monday AI?
All three draft governance prose well, grounded in data you already hold — Copilot even cites back to the source file. All three also tell you, in writing, that the checking is yours: Microsoft says Copilot may produce ungrounded results, Atlassian makes no warranty as to the accuracy of output, monday supplies its AI as is. PMOpack runs no model in its validation layer at all. The AI drafts the narrative; deterministic code checks the figures.
Why not just use Claude Code to make the slides?
Claude Code can create a review deck or document from repository context — a useful creation workspace. PMOpack is the client-reporting system behind the artefact: it keeps the programme snapshot and source trail, checks the reporting pack, and records the publish decision. It gives an independent advisor one governed base across a client portfolio, rather than a fresh prompt for each deck.
Your board is already sceptical. That is the right instinct.
The trust gap around AI-produced information is measured, and it is widest among the people who use AI most. PMOpack is built for the side of that gap that answers to a board.
trust companies to manage AI-related data effectively. Only 41% trust governments.
would be comfortable relying on fully autonomous, agent-led systems.
sceptical are often the heaviest users — educated, white-collar workers who use AI most.
I asked Microsoft Copilot to build a SteerCo pack. Here’s what it gave me.
On 30 April 2026 I ran the exact request a PM runs every week. Beautifully formatted — and the numbers didn’t reconcile. That gap, between a document that looks right and a record that is right, is the whole reason PMOpack exists.
Built by someone who’s sat in the meetings.
17 years embedded in programme delivery — 35+ programmes, £75M+ documented savings, four countries. I’ve sat in 200+ SteerCo meetings and watched accurate-looking packs fall apart under one sharp question from the board. PMOpack is the tool I wanted in the room. I wrote the integrity rules from the failure modes I kept seeing.
- 17
- years in delivery
- 35+
- programmes
- £75M+
- documented savings
- 4
- countries
Check one pack.
See what your board would ask.
Figures from an uploaded file stay traceable to that file and, where the PM has marked it verified, its sign-off, with every accepted exception on record.
Want a walkthrough first? Talk to us about early access