Project Intelligence reads your project's documents and returns a structured, independent assurance report — the blind spots, the failure patterns, what's missing, and the decisions you need to make — in minutes, not a six-week engagement.
Business case, plan, risks, stakeholders, contracts, architecture, cutover plan.
Against what a complete programme of this type should contain — finding the gaps, not summarizing the docs.
A forwardable assurance report: verdict, blind spots, missing information, recommendations — with every finding traceable.
Pick a size — small, medium, large or enterprise. Each is a realistic, fully-worked example so you can see exactly what the review produces at your scale.
The examples are pre-computed runs — you are viewing a saved analysis. Your own analysis runs live and takes longer.
Upload your real project documents — Word, Excel, PowerPoint, PDF or text. Add a few details and run the same review.
Every review applies the same disciplines used by people who can't afford to be wrong — intelligence tradecraft, aviation safety, the markets, and frontier AI — not generic chatbot output.
The disciplines above aren't branding. Select each one to see what it actually changes about your report.
Most tools hand your documents back to you as a tidy summary. Project Intelligence does the opposite: it works out why the programme is most likely to fail — the missing artifacts, the unguarded dependencies, the assumptions no one has tested — and names the decisions that would change the outcome. Like an accident investigation, the point isn't to recount what's on the page; it's to determine the probable cause before it becomes history.
Not every finding deserves equal weight, and the review never pretends otherwise. Each one is graded by the strength of the evidence behind it, so a hard, document-backed gap is visibly separated from an early signal worth watching. You spend attention where it's warranted — and you can defend every escalation, because the confidence is on the page, not in someone's head.
Teams are rarely the best judges of their own programme — proximity hides the cracks. The review is deliberately adversarial: its job is to challenge the plan, not flatter it, and to surface the blind spots an internal team can't or won't raise about its own work. It's the obligation to dissent, built into software and applied without politics or career risk.
A finding you can't trace is just an opinion. Every conclusion in the report links back to the exact document and passage it came from, so anyone — your sponsor, an auditor, a sceptical board member — can follow the reasoning to its source and check it for themselves. Nothing is asserted on authority; the trail is part of the output.
It would be easy to dress the report up with confident-sounding figures like "70% of programmes like this fail." The review refuses to. Where it has no grounded basis for a number, it says so rather than inventing one, and it tells you plainly when the evidence is too thin to conclude. Calibrated honesty is slower to impress — and far harder to regret.
Every review resolves to one of four decision postures — worst to best.
Confidence is stated per finding and capped by the evidence behind it — there is no global "high." The overall reading preserves the worst signal rather than averaging it away.
Your documents and project memory stay private to your account — see Security & data handling for exactly how your data is treated. The review is independent and states its confidence per finding; it never invents "X% of programmes" benchmark figures it cannot stand behind.
Every review your organization runs is stored here — projects, reports, findings, verdicts, and the predictions the product froze to learn from later.