The catalog

KENNEDY Flash Intelligence Brief

An analyst-grade brief on any entity, region, or event. In hours. Every sentence labeled verified or assessed.

An analyst-grade brief on any entity, region, or event. In hours. Every sentence labeled verified or assessed.

Something just happened and you need to understand it today: a coup rumor in a country your supply chain runs through, a counterparty you are about to wire money to, a port strike, a sanctions designation. KENNEDY Flash Brief takes one named tasking and runs it hard through the same open-source intelligence engine that powers our standing watches, then a human analyst reviews it, and what lands on your desk separates what we can prove (with a source you can open) from what we think (with the reasoning stated). No retainer, no three-week consulting quote.

Task your first brief See a sample brief

THE PROBLEM

The intelligence market has two shapes, and neither fits the day you actually need it. Shape one is the retainer: excellent analysts, standing coverage, priced for companies with an intelligence budget line, which you do not have for a question that comes up four times a year. Shape two is bespoke consulting: also excellent, quoted in weeks, and your decision is on Thursday. So what actually happens on the day is shape three: someone senior spends six hours googling, pastes fragments into a chat window, and gets back a fluent, confident summary that cannot tell them which of its own sentences are sourced and which are made up. Then that summary walks into a decision about money, routing, or people. The failure is not the missing information, most of it is public. The failure is that nobody can tell, afterward, which parts of the picture were real.

HOW IT WORKS

You hand us one tasking: an entity, a region, or an event, and the decision it feeds. The engine is the same ingest-classify-correlate-escalate composition our standing watches run on (armed-conflict event data, vessel and aircraft tracks, official hazard and weather feeds, sanctions lists, public filings, the open press), pointed at your tasking and run once, hard, instead of on a cadence. Sourcing is OSINT only: public sources, no paywall circumvention, nothing classified, no covert collection, and no surveillance of private individuals. The output is built around one structural rule: every claim is either VERIFIED, meaning it is tied to a named public source you can open yourself, receipt attached, or ASSESSED, meaning it is our judgment, with the reasoning and a confidence level stated. Nothing sits between. If the open sources cannot answer part of your question, the brief says exactly that, names what is missing, and does not fill the gap with plausible text. Before anything reaches you, an analyst reviews the brief, and that review is a gate, not a formality: nothing auto-sends.

WHAT YOU GET

One flash brief per tasking, delivered in hours, structured to survive scrutiny: - the bottom line, what we can prove, what we assess, and what remains unknown, up front, - the verified claims, each tied to a named public source with the receipt attached, - the assessed claims, each labeled as judgment, with reasoning and stated confidence, - the declared gaps, what the open sources could not answer, named rather than papered over, - the source ledger, everything consulted, so the brief can be audited later by someone who was not in the room.

WHO IT'S FOR

Supply-chain, security, and corporate-development teams at mid-market firms whose geopolitical questions are episodic, first: real exposure, no analyst on staff. Then law firms, insurers, journalists, and family offices who need a fast, sourced picture of an entity or event. And our own standing-watch customers, for whom any alert can escalate into a full brief. If you need classified access, paywalled proprietary intelligence, covert collection, or a report on a private individual, this is not that, and we will decline the tasking.

PRICING (the ladder)

We price the ladder, not a single number, and every number below is a hypothesis we validate with you, not a commitment. - Single brief, one tasking, one flash brief, priced so the person who owns the question can expense it without a procurement cycle. - Brief pack, a monthly allotment with priority turnaround, for teams whose questions recur. - Watch attach, for Kennedy standing-watch subscribers: any alert the monitor raises can escalate into a full brief at a member rate, with the watch supplying the context the brief starts from. Negotiated with the subscription.

IP is licensed, never assigned. Any step that takes your money is gated and confirmed before it runs, nothing charges silently.

THE PROOF (dogfood)

Our first tasking subject is ourselves. AYA runs on a deliberately narrow dependency surface (one LLM provider, one cloud region, a handful of registrars), and those dependencies sit in real jurisdictions with real weather, real politics, and real chokepoints. The first flash briefs off this line are self-tasked: the same engine, the same verified-vs-assessed discipline, aimed at the question "what threatens the vendors AYA cannot instantly replace." The engine itself is not new for this product: it is the already-built Kennedy standing-watch composition, verified on disk and scored by our pattern quality gate, run in a different duty cycle. We did not write new machinery to sell this, which is exactly the moat: the engine compounds, the tasking changes.

HONEST NOTE

We would rather under-promise, so here is the unvarnished state. The engine exists and is quality-gated, but this product has never shipped a brief: not to an external client, and not yet even to ourselves. The live data feeds behind the richest briefs need credentials we have not yet provisioned, the hours-level turnaround is a design target that has never been clocked against a real tasking, and the analyst review gate, while designed in as a hard stop, has never been exercised against a real delivery. The first paid brief will be a first, and its buyer will know that. What we can promise structurally, because the discipline is built rather than aspirational: no unlabeled claims, no invented certainty, declared gaps, and a human between the engine and your inbox, every time. That refusal to fake certainty is the whole product.

Task your first brief

*This page is a specification. The capability it describes is not built yet, and nothing here is a claim that it runs today.*