Concept prototype
PMO Strategic Command Center
One executive view of a transformation portfolio — from this week's RAG status up to national strategy, and from an optimisation trial up to the decision to adopt it.
All data is illustrative. Altair Airways is a fictional Doha-based carrier created for this concept.
The gap
Executive portfolio reporting is usually assembled by hand. Someone pulls delivery status out of Jira, dates out of MS Project, spend out of the finance system, and pastes it into a slide pack the night before the board meets. By the time leadership sees a RAG status, it is weeks old, disconnected from the numbers underneath it, and impossible to interrogate. Ask "why is this amber?" or "which programme actually moves our strategic target?" and the honest answer is another fortnight of analysis.
The bigger gap is strategy. Most portfolio dashboards stop at delivery. They show what is late and what is over budget, but not what any of it is for. In an organisation working to a national vision, that line — from a workstream this week to a 2030 target — is the whole point, and it is almost never drawn.
There is a third gap, and it is where new technology goes to die: the trial. An organisation tests something promising, the results look encouraging, and then nothing happens — because nobody defined what "good enough" meant, the people who would use it never trusted it, and the decision to adopt it never reached the right table.
So I built a concept that closes all three.
What it does
The portfolio
- A weekly reporting model, not a snapshot. A reporting-week selector (Thursday week-endings, the Qatar working week) drives every figure, and RAG status moves week to week so you see the trend, not just today's colour.
- One view, full drill-down. Seven programmes, each opening into a weekly status report, milestones with baseline-versus-forecast dates, a workstream-by-week plan on a page, and steering committee decisions and actions.
- Attention in one place. Escalations, watch items, decisions due and slipping milestones, consolidated — so the meeting opens on what needs a decision.
- Innovation as a governed portfolio. A stage-gate funnel from ideation to scale, where every initiative opens a real business case: problem, benefit model, investment ask, gate criteria and key risks.
- A line to national strategy. The four pillars of Qatar National Vision 2030, mapped to the measurable targets of the Third National Development Strategy.
The trial
- An optimisation trial, end to end. Aircraft stand allocation at a fictional Doha hub: one operating day, 14 stands, about 50 flights. The current plan — fixed rules plus manual fixes — side by side with an optimised plan, and the difference in plan changes, tows, remote-stand use, on-time departures, delay minutes and planner hours.
- Watch it re-plan. Delay or cancel a flight and the plan recovers in seconds, showing what moved and why.
- Explain every choice. Click any flight to see why it got its stand and what each alternative would have cost. Planners can accept or override — and an override needs a reason.
- Governance with a live number. Every accept and override feeds the acceptance rate, measured against the stage-gate criterion of 80%. It currently sits at 79% over a 21-day shadow period [check] — so the SteerCo decision is genuinely open.
The experience
- A two-minute story for first-time visitors. The problem in plain words, a late flight re-planned in front of you, why planners can trust the result, and the decision. One thing on screen at a time, with every term explained.
- One site, seven viewpoints, one click behind the story. A "view as" bar switches between stand planner, duty manager, ground handling, airline station manager, solution integrator, PMO and executive. Each sees only what it needs; only the planner can change the plan.
- Show data sources. One switch pins a tag to every panel: which system the data comes from, whether a machine or a person maintains it, how fresh it is, and whether that feed exists today, needs integration, or needs a data-sharing agreement.
How it's built
Two self-contained pages — vanilla HTML, CSS and JavaScript, no framework, no backend, no build step — with an automated browser test suite behind them. That was deliberate: the audience is leadership and operations, not engineers, and a tool that opens instantly and reads cleanly makes the point better than a build pipeline would.
The architecture it represents is the real one. The portfolio side is a Power BI model over Jira, MS Project, Oracle ERP, Salesforce and SharePoint. The trial side reads the airport's operational database and stand master data, and calls an optimisation engine in shadow mode — it suggests, it never writes back.
The solver in the prototype is deliberately simple: a greedy heuristic, labelled as such on the screen. The concept is engine-agnostic. The planner's screen, the explain panel and the governance stay the same whatever engine runs underneath — and the trial is designed to prove, fairly, whether that engine wins.
The key design decision is the split between what a machine reports and what a person judges. Dates, spend and plans are system-fed. RAG rationale, decisions and overrides are made by the people accountable for them — and recorded with their reasons. The dashboard is built to carry judgment, not replace it.
The numbers are real. The airline isn't.
Altair Airways is fictional — named for the navigation star whose Arabic name, an-nasr aṭ-ṭā'ir, means "the flying eagle". A concept built on a real company's brand can never be published cleanly, and it narrows the work to one employer. A fictional carrier inside a real strategic framework keeps the piece honest and makes the method transferable.
What is not fictional is the strategy. The Vision 2030 pillars and the national development targets are researched and sourced, not invented — 4% average annual GDP growth, 2% productivity growth, a 46% skilled-worker share of the workforce. A Qatari executive knows these numbers.
The harder discipline was resisting green — twice. On the portfolio, the data has real texture: an escalated red programme, slipped milestones, decisions sitting unmade. On the trial, my first results were too good. A 16-point jump in on-time departures from stand allocation alone is not credible; stands are one cause of delay among many. The recalibrated gains are modest — on-time departures from 74% to 80% [check], tows from four to one [check] — and measured like for like, with both plans facing the same day of disruption. Modest, believable gains earn trust. Dramatic ones lose the room.
Designing for the stranger
My first version showed everything to everyone: twenty controls before the first sentence. It impressed builders and lost newcomers. So the site now tells one story first, in plain words, and keeps the detail one click away. Every release was checked the same way: open it as someone who has never heard of stand allocation, and write down what they understand in the first ten seconds. Designing for the person who has never seen the problem is most of the job.
The honest part
The screen was never the hard problem. Neither is the engine.
A command centre does not fix a PMO. It surfaces what the reporting discipline already produces — and if that discipline is weak, the tool just makes weak reporting prettier and faster. I have seen portfolio tooling fail for exactly these reasons, and never once for a missing feature.
The trial has the same lesson. Turn on "Show data sources" and the real blockers are visible on the screen: flight data exists today; live equipment status needs integration; connection data and ground-handling data need agreements between separate organisations. An optimisation engine is only as good as the data it is allowed to see — and that is a negotiation, not a code problem.
And the best engine still fails if planners don't trust it. That is why the explain panel, the override reasons and the acceptance rate exist. A trial succeeds when the people who would use the tool choose to use it, and the decision to adopt it is made on evidence everyone agreed in advance. That is programme work, not maths.
What I'd do next
- Add operating scenarios to the trial: evening delays, a stand out of service, a peak event day.
- Compare three plans, not two: the current method, a classical solver, and the new engine — the fair test any client will judge an engine by.
- Wire the portfolio to a live Power BI model, with role-based security.
- Run the trial view against one airport's real operational data in shadow mode.
Built by Shiraz Hussein. Planned and built with Claude, shipped with Claude Code, reviewed and tested by hand.