When you ask Pilot5 one question, here is exactly what fires.

Before any AI model reasons, Pilot5 builds the brief with you: the decision, the discriminating context, hard constraints, your working hypothesis, your profile, what a useful answer looks like — plus what it infers silently about perimeter, certainty, biases, and preferred output format. You see what's been captured. You correct it. You approve it. The instruction that reaches the deliberation is one you validated. This eliminates the most common reason AI answers miss the point.

Before the personas reason, multiple systems run simultaneously: a live web-search agent issues entity-enriched queries against the open web, a frame analyst reframes your question to surface the real decision behind what was literally asked, a persona gap analyzer runs five parallel calls asking each perspective what specific data it needs and dispatches targeted retrievals to fill those gaps, and a hybrid RAG layer pulls from 400+ verified institutional sources — IMF, EUR-Lex, PubMed, WHO, OECD, SEC, NICE, ClinicalTrials.gov, arXiv, Cochrane, Eurostat, INSEE, BOFIP, EPO and more — across 18 domains. Every claim the personas generate carries a provenance tag: [SOURCED] when it traces to a verifiable reference, [INFERRED] when it's analytical inference. You always know what is grounded and what is reasoning.

Each perspective does not just analyze your question. It analyzes it at full depth and entirely through the lens of its own mandate. Five complete analyses. Five distinct angles. Not five versions of the same answer.

• The Architect reasons about structure, system dependencies, benchmarks, and operational rigor.
• The Strategist reads macro trends, market signals, competitive dynamics, and timing.
• The Engineer examines technical feasibility, math and code accuracy, and execution constraints.
• The Counsel surfaces ethics, nuanced risk, regulatory exposure, and clarity of obligation.
• The Contrarian sole mandate: challenge every assumption the others make.

Then they challenge each other anonymously, labeled only as ANALYSIS_1 through ANALYSIS_5, so each perspective attacks ideas on merit, not on author. If consensus exceeds 90% after Round 2, the Devil's Advocate triggers automatically to stress-test the convergence before it hardens.

Then they converge. The output: GO / PIVOT / STOP. Confidence score. Decision matrix. Action plan with owners and deadlines. Minority Report. Testable predictions.

5 frontier models. A dozen specialized agents around the five-persona core. 400+ verified sources across 18 domains. All of it for one question. Your question.